


| Method | M-Mm to U |
| Unified sources | 57.1 |
| Best single source | 59.8 |
| DIAL [29] - Unified sources | 81.7 |
| DIAL [29] - Best single source | 81.9 |
| 82.5 | |
| 82.2 | |
| 82.7 | |
| 82.4 | |
| mDA | 82.4 |
| Multi-source DA | 84.2 |
| 方法 | M - 毫米到U |
| 统一源 | 57.1 |
| 最佳单一源 | 59.8 |
| DIAL [29] - 统一源 | 81.7 |
| DIAL [29] - 最佳单一源 | 81.9 |
| 82.5 | |
| 82.2 | |
| 82.7 | |
| 82.4 | |
| mDA | 82.4 |
| 多源领域适应(Multi - source Domain Adaptation) | 84.2 |
| Method | SVHN | MNIST-m | Mean | |
| Unified sources | Source only) | 74.1 | 64.4 | 69.3 |
| Source only | 77.7 | 59.4 | 68.6 | |
| Source only from [286] | 72.2 | 64.1 | 68.2 | |
| RevGrad [77] | 68.9 | 71.6 | 70.3 | |
| DAN [154] | 71.0 | 66.6 | 68.8 | |
| DIAL [29] | 82.2 | 68.8 | 75.5 | |
| 82.4 | 69.1 | 75.8 | ||
| mDA | 82.6 | 70.1 | 76.4 | |
| Multi-source | Only Source [286] | 64.6 | 60.7 | 62.7 |
| RevGRAD [77] | 61.4 | 71.1 | 66.3 | |
| DAN [154] | 62.9 | 62.6 | 62.8 | |
| DCTN [286] | 77.5 | 70.9 | 74.2 | |
| Multi-source DA | 84.1 | 69.4 | 76.8 |
| 方法 | 街景门牌号数据集(Street View House Numbers, SVHN) | MNIST-m数据集 | 均值 | |
| 统一源 | 仅源域) | 74.1 | 64.4 | 69.3 |
| 仅源域 | 77.7 | 59.4 | 68.6 | |
| 仅来自[286]的源域 | 72.2 | 64.1 | 68.2 | |
| 反向梯度(RevGrad) [77] | 68.9 | 71.6 | 70.3 | |
| 深度平均网络(Deep Averaging Network, DAN) [154] | 71.0 | 66.6 | 68.8 | |
| DIAL方法 [29] | 82.2 | 68.8 | 75.5 | |
| 82.4 | 69.1 | 75.8 | ||
| 多层去噪自编码器(Multi-layer Denoising Autoencoder, mDA) | 82.6 | 70.1 | 76.4 | |
| 多源 | 仅源域 [286] | 64.6 | 60.7 | 62.7 |
| 反向梯度(RevGRAD) [77] | 61.4 | 71.1 | 66.3 | |
| 深度平均网络(Deep Averaging Network, DAN) [154] | 62.9 | 62.6 | 62.8 | |
| 深度协同训练网络(Deep Co-Training Network, DCTN) [286] | 77.5 | 70.9 | 74.2 | |
| 多源领域自适应(Multi-source Domain Adaptation, Multi-source DA) | 84.1 | 69.4 | 76.8 |
| Method | Sketch | Photo | Art | Cartoon | Mean |
| ResNet [98] | 60.1 | 92.9 | 74.7 | 72.4 | 75.0 |
| DIAL [29] | 66.8 | 97.0 | 87.3 | 85.5 | 84.2 |
| 69.6 | 97.0 | 87.7 | 86.9 | 85.3 | |
| mDA | 70.7 | 97.0 | 87.4 | 86.3 | 85.4 |
| Multi-source DA | 71.6 | 96.6 | 87.5 | 87.0 | 85.7 |
| 方法 | 草图 | 照片 | 艺术 | 卡通 | 均值 |
| 残差网络(ResNet) [98] | 60.1 | 92.9 | 74.7 | 72.4 | 75.0 |
| DIAL [29] | 66.8 | 97.0 | 87.3 | 85.5 | 84.2 |
| 69.6 | 97.0 | 87.7 | 86.9 | 85.3 | |
| 多重去噪自编码器(mDA) | 70.7 | 97.0 | 87.4 | 86.3 | 85.4 |
| 多源领域自适应(Multi - source DA) | 71.6 | 96.6 | 87.5 | 87.0 | 85.7 |
| Method/Targets | Photo Art | Photo Cartoon | Photo Sketch | Art Cartoon | Art Sketch | Cartoon Sketch | Mean |
| ResNet [98] | 71.4 | 84.2 | 81.4 | 62.2 | 70.3 | 54.2 | 70.6 |
| DIAL [29] | 86.7 | 86.5 | 86.8 | 77.1 | 72.1 | 67.7 | 79.5 |
| Random assignment | 86.6 | 86.7 | 85.9 | 76.2 | 69.1 | 69.4 | 79.1 |
| 86.8 | 86.5 | 86.7 | 78.6 | 73.8 | 68.7 | 80.2 | |
| 82.4 | 85.0 | 83.7 | 71.7 | 74.0 | 68.8 | 76.4 | |
| 86.1 | 87.9 | 87.9 | 79.3 | 79.9 | 74.9 | 82.6 | |
| mDA | 87.2 | 88.1 | 88.7 | 77.7 | 81.3 | 77.0 | 83.3 |
| Multi-source/target DA | 87.7 | 88.9 | 86.8 | 79.0 | 79.8 | 75.6 | 83.0 |
| 方法/目标 | 摄影艺术 | 摄影卡通 | 摄影素描 | 艺术卡通 | 艺术素描 | 卡通素描 | 均值 |
| 残差网络(ResNet) [98] | 71.4 | 84.2 | 81.4 | 62.2 | 70.3 | 54.2 | 70.6 |
| 对话式交互自适应学习(DIAL) [29] | 86.7 | 86.5 | 86.8 | 77.1 | 72.1 | 67.7 | 79.5 |
| 随机分配 | 86.6 | 86.7 | 85.9 | 76.2 | 69.1 | 69.4 | 79.1 |
| 86.8 | 86.5 | 86.7 | 78.6 | 73.8 | 68.7 | 80.2 | |
| 82.4 | 85.0 | 83.7 | 71.7 | 74.0 | 68.8 | 76.4 | |
| 86.1 | 87.9 | 87.9 | 79.3 | 79.9 | 74.9 | 82.6 | |
| 多去噪自编码器(mDA) | 87.2 | 88.1 | 88.7 | 77.7 | 81.3 | 77.0 | 83.3 |
| 多源/目标领域自适应(Multi-source/target DA) | 87.7 | 88.9 | 86.8 | 79.0 | 79.8 | 75.6 | 83.0 |





| Source Method Target | A-W D | A-D W | W-D A | Mean | |
| Best single source [286] | TCA [198] | 95.2 | 93.2 | 51.6 | 68.8 |
| GFK [86] | 95.0 | 95.6 | 52.4 | 68.7 | |
| DDC 260 | 98.5 | 95.0 | 52.2 | 70.7 | |
| DRCN [80] | 99.0 | 96.4 | 56.0 | 73.6 | |
| RevGrad[77] | 99.2 | 96.4 | 53.4 | 74.3 | |
| DAN [154] | 99.0 | 96.0 | 54.0 | 72.9 | |
| RTN [155] | 99.6 | 96.8 | 51.0 | 73.7 | |
| Unified sources | Source only from [286] | 98.1 | 93.2 | 50.2 | 80.5 |
| Source only | 94.6 | 89.1 | 49.1 | 77.6 | |
| Source only | 91.9 | 92.7 | 46.5 | 77.0 | |
| RevGrad[286] | 98.8 | 96.2 | 54.6 | 83.2 | |
| DAN [286] | 98.8 | 95.2 | 53.4 | 82.5 | |
| Single BN | 92.9 | 95.2 | 60.1 | 82.7 | |
| DIAL [29] | 93.8 | 94.3 | 62.5 | 83.5 | |
| 93.7 | 94.6 | 62.6 | 83.6 | ||
| mDA | 93.6 | 93.6 | 62.4 | 83.2 | |
| Multi- source | Source only [286] | 98.2 | 92.7 | 51.6 | 80.8 |
| sFRAME[282] | 54.5 | 52.2 | 32.1 | 46.3 | |
| SGF[91] | 39.0 | 52.0 | 28.0 | 39.7 | |
| DCTN [286] | 99.6 | 96.9 | 54.9 | 83.8 | |
| Multi-source DA | 94.8 | 95.8 | 62.9 | 84.5 |
| 源方法 目标 | A-W D | A-D W | W-D A | 均值 | |
| 最佳单源 [286] | 迁移成分分析(TCA) [198] | 95.2 | 93.2 | 51.6 | 68.8 |
| 高斯场核(GFK) [86] | 95.0 | 95.6 | 52.4 | 68.7 | |
| 深度判别相关(DDC) 260 | 98.5 | 95.0 | 52.2 | 70.7 | |
| 深度残差相关网络(DRCN) [80] | 99.0 | 96.4 | 56.0 | 73.6 | |
| 逆向梯度(RevGrad)[77] | 99.2 | 96.4 | 53.4 | 74.3 | |
| 深度对抗网络(DAN) [154] | 99.0 | 96.0 | 54.0 | 72.9 | |
| 循环转移网络(RTN) [155] | 99.6 | 96.8 | 51.0 | 73.7 | |
| 统一源 | 仅源来自 [286] | 98.1 | 93.2 | 50.2 | 80.5 |
| 仅源 | 94.6 | 89.1 | 49.1 | 77.6 | |
| 仅源 | 91.9 | 92.7 | 46.5 | 77.0 | |
| 逆向梯度(RevGrad)[286] | 98.8 | 96.2 | 54.6 | 83.2 | |
| 深度对抗网络(DAN) [286] | 98.8 | 95.2 | 53.4 | 82.5 | |
| 单批量归一化(Single BN) | 92.9 | 95.2 | 60.1 | 82.7 | |
| 领域自适应集成学习(DIAL) [29] | 93.8 | 94.3 | 62.5 | 83.5 | |
| 93.7 | 94.6 | 62.6 | 83.6 | ||
| 多领域自适应(mDA) | 93.6 | 93.6 | 62.4 | 83.2 | |
| 多源 | 仅源 [286] | 98.2 | 92.7 | 51.6 | 80.8 |
| 单帧自适应(sFRAME)[282] | 54.5 | 52.2 | 32.1 | 46.3 | |
| 结构引导特征(SGF)[91] | 39.0 | 52.0 | 28.0 | 39.7 | |
| 深度因果迁移网络(DCTN) [286] | 99.6 | 96.9 | 54.9 | 83.8 | |
| 多源领域自适应(Multi - source DA) | 94.8 | 95.8 | 62.9 | 84.5 |

| Sources Method Target | A-D W | A-W D | W-D A | Mean |
| Hoffman et al. [104] | 24.8 | 42.7 | 12.8 | 26.8 |
| Xiong et al. [283] | 29.3 | 43.6 | 13.3 | 28.7 |
| Gong et al. (AlexNet) [85] | 91.8 | 94.6 | 48.9 | 78.4 |
| 93.1 | 94.3 | 64.2 | 83.9 | |
| mDA | 94.5 | 94.9 | 64.9 | 84.8 |
| Gopalan et al. [92] | 51.3 | 36.1 | 35.8 | 41.1 |
| Nguyen et al. [190] | 64.5 | 68.6 | 41.8 | 58.3 |
| Lin et al. [148] | 73.2 | 81.3 | 41.1 | 65.2 |
| 源 方法 目标 | A - D W | A - W D | W - D A | 均值 |
| 霍夫曼等人 [104] | 24.8 | 42.7 | 12.8 | 26.8 |
| 熊等人 [283] | 29.3 | 43.6 | 13.3 | 28.7 |
| 龚等人(亚历克斯网络) [85] | 91.8 | 94.6 | 48.9 | 78.4 |
| 93.1 | 94.3 | 64.2 | 83.9 | |
| mDA | 94.5 | 94.9 | 64.9 | 84.8 |
| 戈帕兰等人 [92] | 51.3 | 36.1 | 35.8 | 41.1 |
| 阮等人 [190] | 64.5 | 68.6 | 41.8 | 58.3 |
| 林等人 [148] | 73.2 | 81.3 | 41.1 | 65.2 |
| Source Method Target | A-C W-D | W-D A-C | C-W-D A | Mean |
| Gong et al. [85] - original | 41.7 | 35.8 | 41.0 | 39.5 |
| Hoffman et al. [104] - ensemble | 31.7 | 34.4 | 38.9 | 35.0 |
| Hoffman et al. [104] - matching | 39.6 | 34.0 | 34.6 | 36.1 |
| Gong et al. [85] - ensemble | 38.7 | 35.8 | 42.8 | 39.1 |
| Gong et al. [85] - matching | 42.6 | 35.5 | 44.6 | 40.9 |
| Gong et al. (AlexNet) [85] - ensemble | 87.8 | 87.9 | 93.6 | 89.8 |
| 93.5 | 88.2 | 93.7 | 91.8 | |
| mDA | 95.0 | 88.7 | 93.9 | 92.5 |
| 源方法 目标 | A - C W - D | W - D A - C | C - W - D A | 均值 |
| 龚等人 [85] - 原始方法 | 41.7 | 35.8 | 41.0 | 39.5 |
| 霍夫曼等人 [104] - 集成方法 | 31.7 | 34.4 | 38.9 | 35.0 |
| 霍夫曼等人 [104] - 匹配方法 | 39.6 | 34.0 | 34.6 | 36.1 |
| 龚等人 [85] - 集成方法 | 38.7 | 35.8 | 42.8 | 39.1 |
| 龚等人 [85] - 匹配方法 | 42.6 | 35.5 | 44.6 | 40.9 |
| 龚等人(亚历克斯网络) [85] - 集成方法 | 87.8 | 87.9 | 93.6 | 89.8 |
| 93.5 | 88.2 | 93.7 | 91.8 | |
| 多层去噪自编码器(mDA) | 95.0 | 88.7 | 93.9 | 92.5 |


| Net | Norm. | Fr.C | Fr.N | Fr.S | Lj.C | Lj.N | Lj.S | Sa.C | Sa.N | Sa.S | avg. |
| AlexNet | BN | 97.3 | 89.1 | 97.4 | 92.9 | 64.4 | 94.2 | 75.6 | 69.7 | 44.0 | 80.5 |
| WBN | 98.1 | 91.3 | 97.1 | 93.1 | 65.1 | 94.1 | 77.7 | 68.8 | 50.2 | 81.7 | |
| WBN* | 97.1 | 91.9 | 98.0 | 93.9 | 65.6 | 95.0 | 77.2 | 69.9 | 49.9 | 82.1 | |
| ResNet | BN | 97.7 | 82.2 | 90.7 | 89.5 | 61.2 | 90.3 | 70.7 | 73.0 | 38.7 | 77.1 |
| WBN | 98.1 | 81.8 | 94.1 | 94.5 | 61.7 | 93.7 | 75.8 | 76.9 | 37.8 | 79.4 | |
| WBN* | 97.9 | 81.3 | 93.4 | 94.7 | 65.1 | 94.6 | 78.1 | 76.5 | 38.5 | 80.0 |
| 网络 | 范数 | Fr.C | Fr.N | Fr.S | Lj.C | Lj.N | Lj.S | Sa.C | Sa.N | Sa.S | 平均值 |
| 亚历克斯网络(AlexNet) | 批量归一化(BN) | 97.3 | 89.1 | 97.4 | 92.9 | 64.4 | 94.2 | 75.6 | 69.7 | 44.0 | 80.5 |
| 加权批量归一化(WBN) | 98.1 | 91.3 | 97.1 | 93.1 | 65.1 | 94.1 | 77.7 | 68.8 | 50.2 | 81.7 | |
| 加权批量归一化*(WBN*) | 97.1 | 91.9 | 98.0 | 93.9 | 65.6 | 95.0 | 77.2 | 69.9 | 49.9 | 82.1 | |
| 残差网络(ResNet) | 批量归一化(BN) | 97.7 | 82.2 | 90.7 | 89.5 | 61.2 | 90.3 | 70.7 | 73.0 | 38.7 | 77.1 |
| 加权批量归一化(WBN) | 98.1 | 81.8 | 94.1 | 94.5 | 61.7 | 93.7 | 75.8 | 76.9 | 37.8 | 79.4 | |
| 加权批量归一化*(WBN*) | 97.9 | 81.3 | 93.4 | 94.7 | 65.1 | 94.6 | 78.1 | 76.5 | 38.5 | 80.0 |


| Net | Norm. | Fr.C | Sa.C | Lj.C | Fr.N | Sa.N | Lj.N | Fr.S | Lj.S | Sa.S | avg. |
| AlexNet | BN | 26.0 | 38.4 | 34.4 | 27.9 | 26.6 | 33.1 | 28.8 | 34.2 | 25.1 | 30.5 |
| WBN | 25.8 | 38.2 | 33.0 | 29.4 | 26.6 | 34.8 | 30.3 | 36.9 | 25.1 | 31.1 | |
| WBN* | 25.9 | 40.3 | 33.4 | 28.0 | 27.6 | 34.9 | 31.5 | 44.3 | 28.6 | 32.7 | |
| ResNet | BN | 37.9 | 40.9 | 39.3 | 30.8 | 48.3 | 41.2 | 30.6 | 40.6 | 27.6 | 37.5 |
| WBN | 37.3 | 39.5 | 42.6 | 40.4 | 51.8 | 41.0 | 33.8 | 39.6 | 30.8 | 39.6 | |
| WBN* | 36.6 | 40.3 | 40.0 | 41.2 | 56.2 | 45.2 | 35.4 | 39.4 | 25.6 | 40.0 |
| 网络 | 范数(Norm.) | Fr.C | Sa.C | Lj.C | Fr.N | Sa.N | Lj.N | Fr.S | Lj.S | Sa.S | 平均值(avg.) |
| 亚历克斯网络(AlexNet) | 批量归一化(BN) | 26.0 | 38.4 | 34.4 | 27.9 | 26.6 | 33.1 | 28.8 | 34.2 | 25.1 | 30.5 |
| WBN | 25.8 | 38.2 | 33.0 | 29.4 | 26.6 | 34.8 | 30.3 | 36.9 | 25.1 | 31.1 | |
| WBN* | 25.9 | 40.3 | 33.4 | 28.0 | 27.6 | 34.9 | 31.5 | 44.3 | 28.6 | 32.7 | |
| 残差网络(ResNet) | 批量归一化(BN) | 37.9 | 40.9 | 39.3 | 30.8 | 48.3 | 41.2 | 30.6 | 40.6 | 27.6 | 37.5 |
| WBN | 37.3 | 39.5 | 42.6 | 40.4 | 51.8 | 41.0 | 33.8 | 39.6 | 30.8 | 39.6 | |
| WBN* | 36.6 | 40.3 | 40.0 | 41.2 | 56.2 | 45.2 | 35.4 | 39.4 | 25.6 | 40.0 |
| Net | H1 | H2 | H3 | H4 | H5 | H6 | avg. |
| AlexNet | 49.8 | 53.4 | 49.2 | 64.4 | 41.0 | 43.4 | 50.2 |
| AlexNet + BN | 54.5 | 54.6 | 55.6 | 69.7 | 41.8 | 45.9 | 53.7 |
| AlexNet + WBN | 54.7 | 51.9 | 61.8 | 70.6 | 43.9 | 46.5 | 54.9 |
| AlexNet + WBN* | 53.5 | 54.6 | 55.7 | 68.1 | 44.3 | 49.9 | 54.3 |
| ResNet | 55.8 | 47.4 | 64.0 | 69.9 | 42.8 | 50.4 | 55.0 |
| ResNet + WBN | 55.7 | 49.5 | 64.7 | 70.2 | 42.1 | 52.0 | 55.7 |
| ResNet + WBN* | 56.8 | 50.9 | 64.1 | 69.3 | 45.1 | 51.6 | 56.5 |
| 网络 | H1 | H2 | H3 | H4 | H5 | H6 | 平均值 |
| 亚历克斯网络(AlexNet) | 49.8 | 53.4 | 49.2 | 64.4 | 41.0 | 43.4 | 50.2 |
| 亚历克斯网络 + 批量归一化(AlexNet + BN) | 54.5 | 54.6 | 55.6 | 69.7 | 41.8 | 45.9 | 53.7 |
| 亚历克斯网络 + 加权批量归一化(AlexNet + WBN) | 54.7 | 51.9 | 61.8 | 70.6 | 43.9 | 46.5 | 54.9 |
| 亚历克斯网络 + 改进的加权批量归一化(AlexNet + WBN*) | 53.5 | 54.6 | 55.7 | 68.1 | 44.3 | 49.9 | 54.3 |
| 残差网络(ResNet) | 55.8 | 47.4 | 64.0 | 69.9 | 42.8 | 50.4 | 55.0 |
| 残差网络 + 加权批量归一化(ResNet + WBN) | 55.7 | 49.5 | 64.7 | 70.2 | 42.1 | 52.0 | 55.7 |
| 残差网络 + 改进的加权批量归一化(ResNet + WBN*) | 56.8 | 50.9 | 64.1 | 69.3 | 45.1 | 51.6 | 56.5 |
| Method | [273] | [65] | [290] | AlexNet | ResNet | ||||
| Config. | SIFT | CE | BF | - | - | WBN* | BN | WBN* | |
| Acc. | 35.0 | 41.9 | 45.6 | 45.9 | 50.0 | 50.253.7 | 54.3 | 55.0 | 56.5 |
| 方法 | [273] | [65] | [290] | 亚历克斯网络(AlexNet) | 残差网络(ResNet) | ||||
| 配置(Config.) | 尺度不变特征变换(SIFT) | 交叉熵(CE) | 暴力匹配(BF) | - | - | 加权批量归一化(WBN*) | 批量归一化(BN) | 加权批量归一化(WBN*) | |
| 准确率(Acc.) | 35.0 | 41.9 | 45.6 | 45.9 | 50.0 | 50.253.7 | 54.3 | 55.0 | 56.5 |
| Net | AMOSNet | AlexNet | ||||
| Config. | Base | BN | WBN | Base | BN | WBN |
| February-to-August | 83.7 | 88.8 | 90.3 | 83.6 | 88.9 | 90.5 |
| August-to-February | 71.2 | 82.7 | 86.1 | 73.9 | 83.1 | 87.0 |
| 网络 | 阿莫斯网络(AMOSNet) | 亚历克斯网络(AlexNet) | ||||
| 配置(Config.) | 基础 | 批量归一化(BN) | 加权批量归一化(WBN) | 基础 | 批量归一化(BN) | 加权批量归一化(WBN) |
| 2月至8月 | 83.7 | 88.8 | 90.3 | 83.6 | 88.9 | 90.5 |
| 8月至2月 | 71.2 | 82.7 | 86.1 | 73.9 | 83.1 | 87.0 |


| Method | 0 | 15 | 30 | 45 | 60 | 75 | Mean |
| CAE 221 | 72.1 | 95.3 | 92.6 | 81.5 | 92.7 | 79.3 | 85.5 |
| MTAE [79] | 82.5 | 96.3 | 93.4 | 78.6 | 94.2 | 80.5 | 87.5 |
| CCSA [185] | 84.6 | 95.6 | 94.6 | 82.9 | 94.8 | 82.1 | 89.1 |
| BSF | 85.6 | 95.0 | 95.6 | 95.5 | 95.9 | 84.3 | 92.0 |
| 方法 | 0 | 15 | 30 | 45 | 60 | 75 | 均值 |
| CAE 221 | 72.1 | 95.3 | 92.6 | 81.5 | 92.7 | 79.3 | 85.5 |
| MTAE [79] | 82.5 | 96.3 | 93.4 | 78.6 | 94.2 | 80.5 | 87.5 |
| CCSA [185] | 84.6 | 95.6 | 94.6 | 82.9 | 94.8 | 82.1 | 89.1 |
| BSF | 85.6 | 95.0 | 95.6 | 95.5 | 95.9 | 84.3 | 92.0 |

| Model | Art | Cartoon | Photo | Sketch | Mean |
| MTAE [79] | 60.3 | 58.7 | 91.1 | 47.9 | 64.5 |
| LRE-SVM [288] | 59.7 | 52.8 | 85.5 | 37.9 | 59.0 |
| uDICA [186] | 64.6 | 64.5 | 91.8 | 51.1 | 68.0 |
| TF-CNN[133] (no ft) | 62.7 | 52.7 | 88.8 | 52.2 | 64.1 |
| TF-CNN[133] | 62.9 | 67.0 | 89.5 | 57.5 | 69.2 |
| MLDG [144] | 66.2 | 66.9 | 88.0 | 59.0 | 70.0 |
| BSF (no ft) | 64.1 | 60.6 | 90.4 | 49.4 | 66.1 |
| BSF | 64.1 | 66.8 | 90.2 | 60.1 | 70.3 |
| AlexNet [133] | 63.3 | 63.1 | 87.7 | 54.1 | 67.1 |
| 模型 | 艺术 | 卡通 | 照片 | 素描 | 均值 |
| 多任务注意力嵌入模型(MTAE) [79] | 60.3 | 58.7 | 91.1 | 47.9 | 64.5 |
| 局部相对熵支持向量机(LRE - SVM) [288] | 59.7 | 52.8 | 85.5 | 37.9 | 59.0 |
| 无监督深度独立分量分析(uDICA) [186] | 64.6 | 64.5 | 91.8 | 51.1 | 68.0 |
| 迁移特征卷积神经网络(TF - CNN)[133](无微调) | 62.7 | 52.7 | 88.8 | 52.2 | 64.1 |
| 迁移特征卷积神经网络(TF - CNN)[133] | 62.9 | 67.0 | 89.5 | 57.5 | 69.2 |
| 多任务学习域泛化(MLDG) [144] | 66.2 | 66.9 | 88.0 | 59.0 | 70.0 |
| 双尺度特征(BSF)(无微调) | 64.1 | 60.6 | 90.4 | 49.4 | 66.1 |
| 双尺度特征(BSF) | 64.1 | 66.8 | 90.2 | 60.1 | 70.3 |
| 亚历克斯网络(AlexNet) [133] | 63.3 | 63.1 | 87.7 | 54.1 | 67.1 |
| Art | Cartoon | Photo | Sketch | |
| 0 | 65.2 | 54.5 | 90.7 | 52.4 |
| 0.25 | 64.1 | 60.6 | 90.4 | 49.4 |
| 0.5 | 63.8 | 61.0 | 90.4 | 49.1 |
| 0.75 | 64.0 | 60.9 | 90.5 | 47.8 |
| 1 | 63.0 | 60.1 | 90.5 | 47.5 |
| 艺术 | 卡通 | 照片 | 素描 | |
| 0 | 65.2 | 54.5 | 90.7 | 52.4 |
| 0.25 | 64.1 | 60.6 | 90.4 | 49.4 |
| 0.5 | 63.8 | 61.0 | 90.4 | 49.1 |
| 0.75 | 64.0 | 60.9 | 90.5 | 47.8 |
| 1 | 63.0 | 60.1 | 90.5 | 47.5 |


| Camera Type | Illumination | ||
| Artificial | Cloudy | Directed | |
| Kinect | | | |
| Webcam | | | |
| 相机类型 | 照明条件 | ||
| 人工照明 | 阴天光照 | 定向照明 | |
| Kinect(体感设备) | | | |
| 网络摄像头 | | | |






| Method | Across Decades | Across Regions |
| Baseline | 82.3 | 89.2 |
| AdaGraph BN | 86.3 | 91.6 |
| AdaGraph SB | 86.0 | 90.5 |
| AdaGraph Full | 87.0 | 91.0 |
| Baseline + Refinement | 86.2 | 91.3 |
| AdaGraph + Refinement | 88.6 | 91.9 |
| DA upper bound | 89.1 | 92.1 |
| 方法 | 跨数十年 | 跨地区 |
| 基线 | 82.3 | 89.2 |
| 自适应图批量归一化(AdaGraph BN) | 86.3 | 91.6 |
| 自适应图子批量归一化(AdaGraph SB) | 86.0 | 90.5 |
| 自适应图全量归一化(AdaGraph Full) | 87.0 | 91.0 |
| 基线 + 细化 | 86.2 | 91.3 |
| 自适应图 + 细化 | 88.6 | 91.9 |
| 领域自适应上限(DA upper bound) | 89.1 | 92.1 |
| Method | Avg. Accuracy |
| Baseline | 54.0 |
| Baseline | 56.1 |
| MRG-Direct | 58.1 |
| MRG-Indirect | 58.2 |
| AdaGraph (metadata) | 60.1 |
| AdaGraph (images) | 60.8 |
| Baseline + Refinement | 59.5 |
| AdaGraph + Refinement | 60.9 |
| DA upper bound | 60.9 |
| 方法 | 平均准确率 |
| 基线 | 54.0 |
| 基线 | 56.1 |
| MRG直接法 | 58.1 |
| MRG间接法 | 58.2 |
| 自适应图(元数据) | 60.1 |
| 自适应图(图像) | 60.8 |
| 基线 + 细化 | 59.5 |
| 自适应图 + 细化 | 60.9 |
| 域适应上限 | 60.9 |
| Method | Accuracy |
| Baseline SVM [139] | 39.7 |
| Baseline + BN | 43.7 |
| CMA+GFK [103] | 43.0 |
| CMA+SA [103] | 42.7 |
| LLRESVM [139] | 43.6 |
| LLRESVM+EDA[139] | 44.3 |
| ONDA (Baseline+Refinement Stats) [166] | 46.5 |
| Baseline + Refinement Full | 47.3 |
| 方法 | 准确率 |
| 基线支持向量机(Baseline SVM) [139] | 39.7 |
| 基线 + 批量归一化(Baseline + BN) | 43.7 |
| 协方差匹配自适应+全局核适配(CMA+GFK) [103] | 43.0 |
| 协方差匹配自适应+子空间对齐(CMA+SA) [103] | 42.7 |
| 局部线性表示支持向量机(LLRESVM) [139] | 43.6 |
| 局部线性表示支持向量机+增强数据扩充(LLRESVM+EDA)[139] | 44.3 |
| 在线无监督域自适应(ONDA,Baseline+Refinement Stats) [166] | 46.5 |
| 基线 + 完全细化(Baseline + Refinement Full) | 47.3 |
| Method | Baseline | Refinement Stats[166] | Refinement Full |
| Accuracy | 81.9 | 87.3 | 88.1 |
| 方法 | 基线 | 细化统计[166] | 完全细化 |
| 准确率 | 81.9 | 87.3 | 88.1 |


| Dataset | Classifier Only [160] | PackNet[160] | Piggyback | BA2 [17] | BAT | Individual [160] | |||
| ↓ | ↑ | [160] | BN | Simple | Full | ||||
| #Params | 1 | 1.10 | 1.16 | 1.17 | 1.03 | 1.17 | 1.17 | 6 | |
| ImageNet | 76.2 | 75.7 | 75.7 | 76.2 | 76.2 | 76.2 | 76.2 | 76.2 | 76.2 |
| CUBS | 70.7 | 80.4 | 71.4 | 80.4 | 82.1 | 81.2 | 82.6 | 82.4 | 82.8 |
| Stanford Cars | 52.8 | 86.1 | 80.0 | 88.1 | 90.6 | 92.1 | 91.5 | 91.4 | 91.8 |
| Flowers | 86.0 | 93.0 | 90.6 | 93.5 | 95.2 | 95.7 | 96.5 | 96.7 | 96.6 |
| WikiArt | 55.6 | 69.4 | 70.3 | 73.4 | 74.1 | 72.3 | 74.8 | 75.3 | 75.6 |
| Sketch | 50.9 | 76.2 | 78.7 | 79.4 | 79.4 | 79.3 | 80.2 | 80.2 | 80.8 |
| Score | 533 | 732 | 620 | 934 | 1184 | 1265 | 1430 | 1458 | 1500 |
| Score/Params | 533 | 665 | 534 | 805 | 1012 | 1228 | 1222 | 1246 | 250 |
| 数据集 | 仅分类器 [160] | PackNet(打包网络)[160] | 搭载式(Piggyback) | BA2 [17] | BAT(可能为特定缩写,需结合上下文确定准确含义) | 单独的 [160] | |||
| ↓ | ↑ | [160] | 批量归一化(Batch Normalization,BN) | 简单的 | 完整的 | ||||
| 参数数量(#Params) | 1 | 1.10 | 1.16 | 1.17 | 1.03 | 1.17 | 1.17 | 6 | |
| ImageNet(图像网络) | 76.2 | 75.7 | 75.7 | 76.2 | 76.2 | 76.2 | 76.2 | 76.2 | 76.2 |
| CUBS(可能为特定数据集名称,需结合上下文确定准确含义) | 70.7 | 80.4 | 71.4 | 80.4 | 82.1 | 81.2 | 82.6 | 82.4 | 82.8 |
| 斯坦福汽车数据集(Stanford Cars) | 52.8 | 86.1 | 80.0 | 88.1 | 90.6 | 92.1 | 91.5 | 91.4 | 91.8 |
| 花卉数据集 | 86.0 | 93.0 | 90.6 | 93.5 | 95.2 | 95.7 | 96.5 | 96.7 | 96.6 |
| WikiArt(维基艺术) | 55.6 | 69.4 | 70.3 | 73.4 | 74.1 | 72.3 | 74.8 | 75.3 | 75.6 |
| 草图 | 50.9 | 76.2 | 78.7 | 79.4 | 79.4 | 79.3 | 80.2 | 80.2 | 80.8 |
| 得分 | 533 | 732 | 620 | 934 | 1184 | 1265 | 1430 | 1458 | 1500 |
| 得分/参数 | 533 | 665 | 534 | 805 | 1012 | 1228 | 1222 | 1246 | 250 |
| Dataset | Classifier | PackNet [160] | Piggyback | BA2 [17] | BAT | Individual [160] | |||
| Only [160] | ↓ | ↑ | [160] | BN | Simple | Full | |||
| #Params | 1 | 1.11 | 1.15 | 1.21 | 1.17 | 1.21 | 1.21 | 6 | |
| ImageNet | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 |
| CUBS | 73.5 | 80.7 | 69.6 | 79.7 | 81.4 | 82.4 | 81.5 | 81.7 | 81.9 |
| Stanford Cars | 56.8 | 84.7 | 77.9 | 87.2 | 90.1 | 92.9 | 91.7 | 91.6 | 91.4 |
| Flowers | 83.4 | 91.1 | 91.5 | 94.3 | 95.5 | 96.0 | 96.7 | 96.9 | 96.5 |
| WikiArt | 54.9 | 66.3 | 69.2 | 72.0 | 73.9 | 71.5 | 75.5 | 75.7 | 76.4 |
| Sketch | 53.1 | 74.7 | 78.9 | 80.0 | 79.1 | 79.9 | 79.9 | 79.8 | 80.5 |
| Score | 324 | 685 | 607 | 946 | 1209 | 1434 | 1506 | 1534 | 1500 |
| Score/Params | 324 | 617 | 547 | 822 | 999 | 1226 | 1245 | 1268 | 250 |
| 数据集 | 分类器 | PackNet [160] | 搭载式(Piggyback) | BA2 [17] | BAT | 单独式 [160] | |||
| 仅 [160] | ↓ | ↑ | [160] | 批量归一化(Batch Normalization,BN) | 简单 | 完整 | |||
| 参数数量(#Params) | 1 | 1.11 | 1.15 | 1.21 | 1.17 | 1.21 | 1.21 | 6 | |
| 图像网(ImageNet) | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 | 74.4 |
| 加州大学圣地亚哥分校鸟类数据集(CUBS) | 73.5 | 80.7 | 69.6 | 79.7 | 81.4 | 82.4 | 81.5 | 81.7 | 81.9 |
| 斯坦福汽车数据集(Stanford Cars) | 56.8 | 84.7 | 77.9 | 87.2 | 90.1 | 92.9 | 91.7 | 91.6 | 91.4 |
| 花卉数据集 | 83.4 | 91.1 | 91.5 | 94.3 | 95.5 | 96.0 | 96.7 | 96.9 | 96.5 |
| 维基艺术数据集(WikiArt) | 54.9 | 66.3 | 69.2 | 72.0 | 73.9 | 71.5 | 75.5 | 75.7 | 76.4 |
| 草图数据集 | 53.1 | 74.7 | 78.9 | 80.0 | 79.1 | 79.9 | 79.9 | 79.8 | 80.5 |
| 分数 | 324 | 685 | 607 | 946 | 1209 | 1434 | 1506 | 1534 | 1500 |
| 分数/参数 | 324 | 617 | 547 | 822 | 999 | 1226 | 1245 | 1268 | 250 |
| Dataset | Classifier Only [160] | PackNet [160] | Piggyback [160] | BAT | Individual [160] | ||
| ↓ | ↑ | Simple | Full | ||||
| #Params | 1 | 1.09 | 1.16 | 1.16 | 1.16 | 6 | |
| ImageNet | 71.6 | 70.7 | 70.7 | 71.6 | 71.6 | 71.6 | 71.6 |
| CUBS | 63.5 | 77.7 | 70.3 | 77.8 | 77.4 | 77.4 | 77.4 |
| Stanford Cars | 45.3 | 84.2 | 78.3 | 86.1 | 87.2 | 87.3 | 87.0 |
| Flowers | 80.6 | 89.7 | 89.8 | 90.7 | 91.6 | 91.5 | 92.3 |
| WikiArt | 50.5 | 67.2 | 68.5 | 71.2 | 71.6 | 71.9 | 67.7 |
| Sketch | 41.5 | 71.4 | 75.1 | 76.5 | 76.5 | 76.7 | 76.4 |
| Score | 342 | 1152 | 979 | 1441 | 1530 | 1538 | 1500 |
| Score/Params | 342 | 1057 | 898 | 1243 | 1319 | 1326 | 250 |
| 数据集 | 仅分类器 [160] | PackNet [160] | 背负式(Piggyback) [160] | 蝙蝠(BAT) | 单独的 [160] | ||
| ↓ | ↑ | 简单的 | 完整的 | ||||
| 参数数量 | 1 | 1.09 | 1.16 | 1.16 | 1.16 | 6 | |
| ImageNet(图像网) | 71.6 | 70.7 | 70.7 | 71.6 | 71.6 | 71.6 | 71.6 |
| CUBS(加州大学伯克利分校鸟类数据集) | 63.5 | 77.7 | 70.3 | 77.8 | 77.4 | 77.4 | 77.4 |
| 斯坦福汽车数据集 | 45.3 | 84.2 | 78.3 | 86.1 | 87.2 | 87.3 | 87.0 |
| 花卉数据集 | 80.6 | 89.7 | 89.8 | 90.7 | 91.6 | 91.5 | 92.3 |
| WikiArt(维基艺术数据集) | 50.5 | 67.2 | 68.5 | 71.2 | 71.6 | 71.9 | 67.7 |
| 草图数据集 | 41.5 | 71.4 | 75.1 | 76.5 | 76.5 | 76.7 | 76.4 |
| 分数 | 342 | 1152 | 979 | 1441 | 1530 | 1538 | 1500 |
| 分数/参数 | 342 | 1057 | 898 | 1243 | 1319 | 1326 | 250 |
| Method | #Par | ImN | Airc | C100 | DP | DTD | GTS | Flw | Ogl | SVHN | UCF | Score | |
| Feature [214] | 1 | 59.7 | 23.3 | 63.1 | 80.3 | 45.4 | 68.2 | 73.7 | 58.8 | 43.5 | 26.8 | 544 | 544 |
| Fine-tune [214] | 10 | 59.9 | 60.3 | 82.1 | 92.8 | 55.5 | 97.5 | 81.4 | 87.7 | 96.6 | 51.2 | 2500 | 250 |
| RA [214] | 2 | 59.7 | 56.7 | 81.2 | 93.9 | 50.9 | 97.1 | 66.2 | 89.6 | 96.1 | 47.5 | 2118 | 1059 |
| RA-decay[214] | 2 | 59.7 | 61.9 | 81.2 | 93.9 | 57.1 | 97.6 | 81.7 | 89.6 | 96.1 | 50.1 | 2621 | 1311 |
| RA-N [215] | 2 | 60.3 | 61.9 | 81.2 | 93.9 | 57.1 | 99.3 | 81.7 | 89.6 | 96.6 | 50.1 | 3159 | 1580 |
| DAN [223] | 2.17 | 57.7 | 64.1 | 80.1 | 91.3 | 56.5 | 98.5 | 86.1 | 89.7 | 96.8 | 49.4 | 2852 | 1314 |
| PA [215] | 2 | 60.3 | 64.2 | 81.9 | 94.7 | 58.8 | 99.4 | 84.7 | 89.2 | 96.5 | 50.9 | 3412 | 1706 |
| MTAN [150] | 1.74 | 63.9 | 61.8 | 81.6 | 91.6 | 56.4 | 98.8 | 81.0 | 89.8 | 96.9 | 50.6 | 2941 | 1690 |
| SpotTune [96] | 11 | 60.3 | 63.9 | 80.5 | 96.5 | 57.1 | 99.5 | 85.2 | 88.8 | 96.7 | 52.3 | 3612 | 328 |
| CovNorm [140] | 1.25 | 60.4 | 69.4 | 81.3 | 98.8 | 60.0 | 99.1 | 83.4 | 87.7 | 96.6 | 48.9 | 3713 | 2970 |
| PB [160] | 1.28 | 57.7 | 65.3 | 79.9 | 97.0 | 57.5 | 97.3 | 79.1 | 87.6 | 97.2 | 47.5 | 2838 | 2217 |
| PB ours | 1.28 | 60.8 | 52.3 | 80.0 | 95.1 | 59.6 | 98.7 | 82.9 | 85.1 | 96.7 | 46.9 | 2805 | 2191 |
| 1.03 | 56.9 | 49.4 | 78.1 | 95.5 | 55.1 | 99.4 | 86.1 | 88.7 | 96.9 | 50.2 | 3199 | 3106 | |
| BAT (S) [171] | 1.29 | 60.8 | 51.3 | 81.9 | 94.7 | 59.0 | 99.1 | 88.0 | 89.3 | 96.5 | 48.7 | 3263 | 2529 |
| BAT (F) | 1.29 | 60.8 | 52.8 | 82.0 | 96.2 | 58.7 | 99.2 | 88.2 | 89.2 | 96.8 | 48.6 | 3497 | 2711 |
| PA-SVD[215] | 1.5 | 60.3 | 66.0 | 81.9 | 94.2 | 57.8 | 99.2 | 85.7 | 89.3 | 96.6 | 52.5 | 3398 | 2265 |
| RA-joint[214] | 2 | 59.2 | 63.7 | 81.3 | 93.3 | 57.0 | 97.5 | 83.4 | 89.8 | 96.2 | 50.3 | 2643 | 1322 |
| 方法 | #段落 | 虚部(ImN) | 气室(Airc) | C100 | 动态规划(DP) | 文档类型定义(DTD) | 全局时间同步(GTS) | 流量(Flw) | 油藏地质力学(Ogl) | 街景门牌号数据集(SVHN) | 佛罗里达大学数据集(UCF) | 分数 | |
| 特征 [214] | 1 | 59.7 | 23.3 | 63.1 | 80.3 | 45.4 | 68.2 | 73.7 | 58.8 | 43.5 | 26.8 | 544 | 544 |
| 微调 [214] | 10 | 59.9 | 60.3 | 82.1 | 92.8 | 55.5 | 97.5 | 81.4 | 87.7 | 96.6 | 51.2 | 2500 | 250 |
| 随机调整(RA) [214] | 2 | 59.7 | 56.7 | 81.2 | 93.9 | 50.9 | 97.1 | 66.2 | 89.6 | 96.1 | 47.5 | 2118 | 1059 |
| 随机调整衰减(RA - decay)[214] | 2 | 59.7 | 61.9 | 81.2 | 93.9 | 57.1 | 97.6 | 81.7 | 89.6 | 96.1 | 50.1 | 2621 | 1311 |
| 随机调整 - N(RA - N) [215] | 2 | 60.3 | 61.9 | 81.2 | 93.9 | 57.1 | 99.3 | 81.7 | 89.6 | 96.6 | 50.1 | 3159 | 1580 |
| 深度对抗网络(DAN) [223] | 2.17 | 57.7 | 64.1 | 80.1 | 91.3 | 56.5 | 98.5 | 86.1 | 89.7 | 96.8 | 49.4 | 2852 | 1314 |
| 投影调整(PA) [215] | 2 | 60.3 | 64.2 | 81.9 | 94.7 | 58.8 | 99.4 | 84.7 | 89.2 | 96.5 | 50.9 | 3412 | 1706 |
| 多任务注意力网络(MTAN) [150] | 1.74 | 63.9 | 61.8 | 81.6 | 91.6 | 56.4 | 98.8 | 81.0 | 89.8 | 96.9 | 50.6 | 2941 | 1690 |
| 局部微调(SpotTune) [96] | 11 | 60.3 | 63.9 | 80.5 | 96.5 | 57.1 | 99.5 | 85.2 | 88.8 | 96.7 | 52.3 | 3612 | 328 |
| 协方差归一化(CovNorm) [140] | 1.25 | 60.4 | 69.4 | 81.3 | 98.8 | 60.0 | 99.1 | 83.4 | 87.7 | 96.6 | 48.9 | 3713 | 2970 |
| 投影块(PB) [160] | 1.28 | 57.7 | 65.3 | 79.9 | 97.0 | 57.5 | 97.3 | 79.1 | 87.6 | 97.2 | 47.5 | 2838 | 2217 |
| 我们的投影块(PB ours) | 1.28 | 60.8 | 52.3 | 80.0 | 95.1 | 59.6 | 98.7 | 82.9 | 85.1 | 96.7 | 46.9 | 2805 | 2191 |
| 1.03 | 56.9 | 49.4 | 78.1 | 95.5 | 55.1 | 99.4 | 86.1 | 88.7 | 96.9 | 50.2 | 3199 | 3106 | |
| 批量对抗训练(S)(BAT (S)) [171] | 1.29 | 60.8 | 51.3 | 81.9 | 94.7 | 59.0 | 99.1 | 88.0 | 89.3 | 96.5 | 48.7 | 3263 | 2529 |
| 批量对抗训练(F)(BAT (F)) | 1.29 | 60.8 | 52.8 | 82.0 | 96.2 | 58.7 | 99.2 | 88.2 | 89.2 | 96.8 | 48.6 | 3497 | 2711 |
| 投影调整 - 奇异值分解(PA - SVD)[215] | 1.5 | 60.3 | 66.0 | 81.9 | 94.2 | 57.8 | 99.2 | 85.7 | 89.3 | 96.6 | 52.5 | 3398 | 2265 |
| 随机调整 - 联合(RA - joint)[214] | 2 | 59.2 | 63.7 | 81.3 | 93.3 | 57.0 | 97.5 | 83.4 | 89.8 | 96.2 | 50.3 | 2643 | 1322 |
| Method | CUBS | CARS | Flowers | WikiArt | Sketch | ||||
| Piggyback [160] | 0 | 0 | 0 | 1 | 80.4 | 88.1 | 93.6 | 73.4 | 79.4 |
| Piggyback* | 0 | 0 | 0 | 1 | 80.4 | 87.8 | 93.1 | 72.5 | 78.6 |
| Piggyback* with BN | 0 | 0 | 0 | 1 | 82.1 | 90.6 | 95.2 | 74.1 | 79.4 |
| Piggyback* with BN | 0 | ✓ | 0 | 1 | 81.9 | 89.9 | 94.8 | 73.7 | 79.9 |
| BAT (Simple, no bias) | 1 | 0 | ✓ | 0 | 80.8 | 90.3 | 96.1 | 73.5 | 80.0 |
| BAT (Simple) [171] | 1 | ✓ | ✓ | 0 | 82.6 | 91.5 | 96.5 | 74.8 | 80.2 |
| BAT (Simple with Sigmoid) | 1 | ✓ | ✓ | 0 | 82.6 | 91.4 | 96.4 | 75.2 | 80.2 |
| BAT (Full, no bias) | 1 | 0 | ✓ | ✓ | 80.7 | 90.2 | 96.0 | 72.0 | 78.8 |
| BAT (Full,no | 1 | ✓ | 0 | ✓ | 80.6 | 87.5 | 91.0 | 73.0 | 78.4 |
| BAT (Full) | 1 | ✓ | ✓ | ✓ | 82.4 | 91.4 | 96.7 | 75.3 | 80.2 |
| BAT (Full with Sigmoid) | 1 | ✓ | ✓ | ✓ | 82.7 | 91.4 | 96.6 | 75.2 | 80.2 |
| BAT (Full, channel-wise) | 1 | ✓ | ✓ | ✓ | 82.0 | 91.0 | 96.3 | 74.8 | 80.0 |
| 方法 | 加州大学圣地亚哥分校鸟类数据集(CUBS) | 斯坦福汽车数据集(CARS) | 花卉数据集 | 维基艺术数据集(WikiArt) | 草图数据集 | ||||
| 搭载式方法(Piggyback) [160] | 0 | 0 | 0 | 1 | 80.4 | 88.1 | 93.6 | 73.4 | 79.4 |
| 改进的搭载式方法(Piggyback*) | 0 | 0 | 0 | 1 | 80.4 | 87.8 | 93.1 | 72.5 | 78.6 |
| 带批量归一化的改进搭载式方法(Piggyback* with BN) | 0 | 0 | 0 | 1 | 82.1 | 90.6 | 95.2 | 74.1 | 79.4 |
| 带批量归一化的改进搭载式方法(Piggyback* with BN) | 0 | ✓ | 0 | 1 | 81.9 | 89.9 | 94.8 | 73.7 | 79.9 |
| 批量自适应训练(简单版,无偏置)(BAT (Simple, no bias)) | 1 | 0 | ✓ | 0 | 80.8 | 90.3 | 96.1 | 73.5 | 80.0 |
| 批量自适应训练(简单版) [171](BAT (Simple) [171]) | 1 | ✓ | ✓ | 0 | 82.6 | 91.5 | 96.5 | 74.8 | 80.2 |
| 带Sigmoid激活函数的批量自适应训练(简单版)(BAT (Simple with Sigmoid)) | 1 | ✓ | ✓ | 0 | 82.6 | 91.4 | 96.4 | 75.2 | 80.2 |
| 批量自适应训练(完整版,无偏置)(BAT (Full, no bias)) | 1 | 0 | ✓ | ✓ | 80.7 | 90.2 | 96.0 | 72.0 | 78.8 |
| 批量自适应训练(完整版,无 | 1 | ✓ | 0 | ✓ | 80.6 | 87.5 | 91.0 | 73.0 | 78.4 |
| 批量自适应训练(完整版)(BAT (Full)) | 1 | ✓ | ✓ | ✓ | 82.4 | 91.4 | 96.7 | 75.3 | 80.2 |
| 带Sigmoid激活函数的批量自适应训练(完整版)(BAT (Full with Sigmoid)) | 1 | ✓ | ✓ | ✓ | 82.7 | 91.4 | 96.6 | 75.2 | 80.2 |
| 按通道的批量自适应训练(完整版)(BAT (Full, channel-wise)) | 1 | ✓ | ✓ | ✓ | 82.0 | 91.0 | 96.3 | 74.8 | 80.0 |
| Method | CUBS | CARS | Flowers | WikiArt | Sketch | ||||
| Piggyback [160] | 0 | 0 | 0 | 1 | 79.7 | 87.2 | 94.3 | 72.0 | 80.0 |
| Piggyback* | 0 | 0 | 0 | 1 | 80.0 | 86.6 | 94.4 | 71.9 | 78.7 |
| Piggyback* with BN | 0 | 0 | 0 | 1 | 81.4 | 90.1 | 95.5 | 73.9 | 79.1 |
| Piggyback* with BN | 0 | ✓ | 0 | 1 | 81.9 | 90.1 | 95.4 | 72.6 | 79.9 |
| BAT (Simple, no bias) | 1 | 0 | ✓ | 0 | 80.4 | 91.4 | 96.7 | 75.0 | 79.7 |
| BAT (Simple) [171] | 1 | ✓ | ✓ | 0 | 81.5 | 91.7 | 96.7 | 75.5 | 79.9 |
| BAT (Simple with Sigmoid) | 1 | ✓ | ✓ | 0 | 81.5 | 91.7 | 97.0 | 76.0 | 79.8 |
| BAT (Full, no bias) | 1 | 0 | ✓ | ✓ | 80.2 | 91.1 | 96.5 | 75.1 | 79.2 |
| BAT (Full,no | 1 | ✓ | 0 | ✓ | 79.8 | 87.2 | 91.8 | 73.2 | 78.1 |
| BAT (Full) [172] | 1 | ✓ | ✓ | ✓ | 81.7 | 91.6 | 96.9 | 75.7 | 79.9 |
| BAT (Full with Sigmoid) | 1 | ✓ | ✓ | ✓ | 82.0 | 91.7 | 97.0 | 76.0 | 79.9 |
| BAT (Full, channel-wise) | 1 | ✓ | ✓ | ✓ | 81.4 | 91.6 | 96.5 | 75.5 | 79.9 |
| 方法 | 加州大学圣地亚哥分校鸟类数据集(CUBS) | 斯坦福汽车数据集(CARS) | 花卉数据集 | 维基艺术数据集(WikiArt) | 草图数据集 | ||||
| 搭载式方法(Piggyback) [160] | 0 | 0 | 0 | 1 | 79.7 | 87.2 | 94.3 | 72.0 | 80.0 |
| 改进的搭载式方法(Piggyback*) | 0 | 0 | 0 | 1 | 80.0 | 86.6 | 94.4 | 71.9 | 78.7 |
| 带批量归一化的改进搭载式方法(Piggyback* with BN) | 0 | 0 | 0 | 1 | 81.4 | 90.1 | 95.5 | 73.9 | 79.1 |
| 带批量归一化的改进搭载式方法(Piggyback* with BN) | 0 | ✓ | 0 | 1 | 81.9 | 90.1 | 95.4 | 72.6 | 79.9 |
| 批量自适应训练(简单版,无偏置)(BAT (Simple, no bias)) | 1 | 0 | ✓ | 0 | 80.4 | 91.4 | 96.7 | 75.0 | 79.7 |
| 批量自适应训练(简单版) [171](BAT (Simple) [171]) | 1 | ✓ | ✓ | 0 | 81.5 | 91.7 | 96.7 | 75.5 | 79.9 |
| 带Sigmoid激活函数的批量自适应训练(简单版)(BAT (Simple with Sigmoid)) | 1 | ✓ | ✓ | 0 | 81.5 | 91.7 | 97.0 | 76.0 | 79.8 |
| 批量自适应训练(完整版,无偏置)(BAT (Full, no bias)) | 1 | 0 | ✓ | ✓ | 80.2 | 91.1 | 96.5 | 75.1 | 79.2 |
| 批量自适应训练(完整版,无 | 1 | ✓ | 0 | ✓ | 79.8 | 87.2 | 91.8 | 73.2 | 78.1 |
| 批量自适应训练(完整版) [172](BAT (Full) [172]) | 1 | ✓ | ✓ | ✓ | 81.7 | 91.6 | 96.9 | 75.7 | 79.9 |
| 带Sigmoid激活函数的批量自适应训练(完整版)(BAT (Full with Sigmoid)) | 1 | ✓ | ✓ | ✓ | 82.0 | 91.7 | 97.0 | 76.0 | 79.9 |
| 按通道的批量自适应训练(完整版)(BAT (Full, channel-wise)) | 1 | ✓ | ✓ | ✓ | 81.4 | 91.6 | 96.5 | 75.5 | 79.9 |




| Method | 19-1 | 15-5 | 15-1 | ||||||
| 1-19 | 20 | all | 1-15 | 16-20 | all | 1-15 | 16-20 | all | |
| FT | 5.8 | 12.3 | 6.2 | 1.1 | 33.6 | 9.2 | 0.2 | 1.8 | 0.6 |
| PI [300] | 5.4 | 14.1 | 5.9 | 1.3 | 34.1 | 9.5 | 0.0 | 1.8 | 0.4 |
| EWC [118] | 23.2 | 16.0 | 22.9 | 26.7 | 37.7 | 29.4 | 0.3 | 4.3 | 1.3 |
| RW [36] | 19.4 | 15.7 | 19.2 | 17.9 | 36.9 | 22.7 | 0.2 | 5.4 | 1.5 |
| LwF [144] | 53.0 | 9.1 | 50.8 | 58.4 | 37.4 | 53.1 | 0.8 | 3.6 | 1.5 |
| LwF-MC [216] | 63.0 | 13.2 | 60.5 | 67.2 | 41.2 | 60.7 | 4.5 | 7.0 | 5.2 |
| ILT [178] | 69.1 | 16.4 | 66.4 | 63.2 | 39.5 | 57.3 | 3.7 | 5.7 | 4.2 |
| MiB | 69.6 | 25.6 | 67.4 | 71.8 | 43.3 | 64.7 | 46.2 | 12.9 | 37.9 |
| Joint | 77.4 | 78.0 | 77.4 | 79.1 | 72.6 | 77.4 | 79.1 | 72.6 | 77.4 |
| 方法 | 19-1 | 15-5 | 15-1 | ||||||
| 1-19 | 20 | 全部 | 1-15 | 16-20 | 全部 | 1-15 | 16-20 | 全部 | |
| 傅里叶变换(FT) | 5.8 | 12.3 | 6.2 | 1.1 | 33.6 | 9.2 | 0.2 | 1.8 | 0.6 |
| 渐进式推理(PI) [300] | 5.4 | 14.1 | 5.9 | 1.3 | 34.1 | 9.5 | 0.0 | 1.8 | 0.4 |
| 弹性权重巩固(EWC) [118] | 23.2 | 16.0 | 22.9 | 26.7 | 37.7 | 29.4 | 0.3 | 4.3 | 1.3 |
| 重放(RW) [36] | 19.4 | 15.7 | 19.2 | 17.9 | 36.9 | 22.7 | 0.2 | 5.4 | 1.5 |
| 学习不遗忘(LwF) [144] | 53.0 | 9.1 | 50.8 | 58.4 | 37.4 | 53.1 | 0.8 | 3.6 | 1.5 |
| 多分类学习不遗忘(LwF - MC) [216] | 63.0 | 13.2 | 60.5 | 67.2 | 41.2 | 60.7 | 4.5 | 7.0 | 5.2 |
| 增量式学习迁移(ILT) [178] | 69.1 | 16.4 | 66.4 | 63.2 | 39.5 | 57.3 | 3.7 | 5.7 | 4.2 |
| 兆字节(MiB) | 69.6 | 25.6 | 67.4 | 71.8 | 43.3 | 64.7 | 46.2 | 12.9 | 37.9 |
| 联合 | 77.4 | 78.0 | 77.4 | 79.1 | 72.6 | 77.4 | 79.1 | 72.6 | 77.4 |
| Method | 19-1 | 15-5 | 15-1 | ||||||
| 1-19 | 20 | all | 1-15 | 16-20 | all | 1-15 | 16-20 | all | |
| FT | 6.8 | 12.9 | 7.1 | 2.1 | 33.1 | 9.8 | 0.2 | 1.8 | 0.6 |
| PI [300] | 7.5 | 14.0 | 7.8 | 1.6 | 33.3 | 9.5 | 0.0 | 1.8 | 0.5 |
| EWC [118] | 26.9 | 14.0 | 26.3 | 24.3 | 35.5 | 27.1 | 0.3 | 4.3 | 1.3 |
| RW [36] | 23.3 | 14.2 | 22.9 | 16.6 | 34.9 | 21.2 | 0.0 | 5.2 | 1.3 |
| LwF [144] | 51.2 | 8.5 | 49.1 | 58.9 | 36.6 | 53.3 | 1.0 | 3.9 | 1.8 |
| LwF-MC [216] | 64.4 | 13.3 | 61.9 | 58.1 | 35.0 | 52.3 | 6.4 | 8.4 | 6.9 |
| ILT [178] | 67.1 | 12.3 | 64.4 | 66.3 | 40.6 | 59.9 | 4.9 | 7.8 | 5.7 |
| MiB | 70.2 | 22.1 | 67.8 | 75.5 | 49.4 | 69.0 | 35.1 | 13.5 | 29.7 |
| Joint | 77.4 | 78.0 | 77.4 | 79.1 | 72.6 | 77.4 | 79.1 | 72.6 | 77.4 |
| 方法 | 19-1 | 15-5 | 15-1 | ||||||
| 1-19 | 20 | 全部 | 1-15 | 16-20 | 全部 | 1-15 | 16-20 | 全部 | |
| 傅里叶变换(FT) | 6.8 | 12.9 | 7.1 | 2.1 | 33.1 | 9.8 | 0.2 | 1.8 | 0.6 |
| 渐进式推理(PI) [300] | 7.5 | 14.0 | 7.8 | 1.6 | 33.3 | 9.5 | 0.0 | 1.8 | 0.5 |
| 弹性权重巩固(EWC) [118] | 26.9 | 14.0 | 26.3 | 24.3 | 35.5 | 27.1 | 0.3 | 4.3 | 1.3 |
| 重放(RW) [36] | 23.3 | 14.2 | 22.9 | 16.6 | 34.9 | 21.2 | 0.0 | 5.2 | 1.3 |
| 学习不遗忘(LwF) [144] | 51.2 | 8.5 | 49.1 | 58.9 | 36.6 | 53.3 | 1.0 | 3.9 | 1.8 |
| 多分类学习不遗忘(LwF - MC) [216] | 64.4 | 13.3 | 61.9 | 58.1 | 35.0 | 52.3 | 6.4 | 8.4 | 6.9 |
| 增量式学习迁移(ILT) [178] | 67.1 | 12.3 | 64.4 | 66.3 | 40.6 | 59.9 | 4.9 | 7.8 | 5.7 |
| 兆字节(MiB) | 70.2 | 22.1 | 67.8 | 75.5 | 49.4 | 69.0 | 35.1 | 13.5 | 29.7 |
| 联合 | 77.4 | 78.0 | 77.4 | 79.1 | 72.6 | 77.4 | 79.1 | 72.6 | 77.4 |
| 19-1 | 15-5 | 15-1 | |||||||
| 1-19 | 20 | all | 1-15 | 16-20 | all | 1-15 | 16-20 | all | |
| LwF [144] | 51.2 | 8.5 | 49.1 | 58.9 | 36.6 | 53.3 | 1.0 | 3.9 | 1.8 |
| 57.6 | 9.9 | 55.2 | 63.2 | 38.1 | 57.0 | 12.0 | 3.7 | 9.9 | |
| 66.0 | 11.9 | 63.3 | 72.9 | 46.3 | 66.3 | 34.8 | 4.5 | 27.2 | |
| +init | 70.2 | 22.1 | 67.8 | 75.5 | 49.4 | 69.0 | 35.1 | 13.5 | 29.7 |
| 19-1 | 15-5 | 15-1 | |||||||
| 1-19 | 20 | 所有 | 1-15 | 16-20 | 所有 | 1-15 | 16-20 | 所有 | |
| 学习不遗忘(Learning without Forgetting,LwF) [144] | 51.2 | 8.5 | 49.1 | 58.9 | 36.6 | 53.3 | 1.0 | 3.9 | 1.8 |
| 57.6 | 9.9 | 55.2 | 63.2 | 38.1 | 57.0 | 12.0 | 3.7 | 9.9 | |
| 66.0 | 11.9 | 63.3 | 72.9 | 46.3 | 66.3 | 34.8 | 4.5 | 27.2 | |
| +初始化 | 70.2 | 22.1 | 67.8 | 75.5 | 49.4 | 69.0 | 35.1 | 13.5 | 29.7 |
| Method | 100-50 | 50-50 | |||||
| 1-100 | 101-150 | all | 1-50 | 51-100 | 101-150 | all | |
| FT | 0.0 | 24.9 | 8.3 | 0.0 | 0.0 | 22.0 | 7.3 |
| LwF [144] | 21.1 | 25.6 | 22.6 | 5.7 | 12.9 | 22.8 | 13.9 |
| LwF-MC [216] | 34.2 | 10.5 | 26.3 | 27.8 | 7.0 | 10.4 | 15.1 |
| ILT [178] | 22.9 | 18.9 | 21.6 | 8.4 | 9.7 | 14.3 | 10.8 |
| MiB | 37.9 | 27.9 | 34.6 | 35.5 | 22.2 | 23.6 | 27.0 |
| Joint | 44.3 | 28.2 | 38.9 | 51.1 | 38.3 | 28.2 | 38.9 |
| 方法 | 100-50 | 50-50 | |||||
| 1-100 | 101-150 | 全部 | 1-50 | 51-100 | 101-150 | 全部 | |
| 微调(Fine-Tuning) | 0.0 | 24.9 | 8.3 | 0.0 | 0.0 | 22.0 | 7.3 |
| 学习不遗忘(Learning without Forgetting) [144] | 21.1 | 25.6 | 22.6 | 5.7 | 12.9 | 22.8 | 13.9 |
| 多分类学习不遗忘(Learning without Forgetting - Multi-Class) [216] | 34.2 | 10.5 | 26.3 | 27.8 | 7.0 | 10.4 | 15.1 |
| 增量式学习训练(Incremental Learning Training) [178] | 22.9 | 18.9 | 21.6 | 8.4 | 9.7 | 14.3 | 10.8 |
| 兆字节(Mebibyte) | 37.9 | 27.9 | 34.6 | 35.5 | 22.2 | 23.6 | 27.0 |
| 联合 | 44.3 | 28.2 | 38.9 | 51.1 | 38.3 | 28.2 | 38.9 |
| Method | 1-100 | 100-110 | 110-120 | 120-130 | 130-140 | 140-150 | all |
| FT | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 16.6 | 1.1 |
| LwF [144] | 0.1 | 0.0 | 0.4 | 2.6 | 4.6 | 16.9 | 1.7 |
| LwF-MC [216] | 18.7 | 2.5 | 8.7 | 4.1 | 6.5 | 5.1 | 14.3 |
| ILT [178] | 0.3 | 0.0 | 1.0 | 2.1 | 4.6 | 10.7 | 1.4 |
| MiB | 31.8 | 10.4 | 14.8 | 12.8 | 13.6 | 18.7 | 25.9 |
| Joint | 44.3 | 26.1 | 42.8 | 26.7 | 28.1 | 17.3 | 38.9 |
| 方法 | 1-100 | 100-110 | 110-120 | 120-130 | 130-140 | 140-150 | 全部 |
| 微调(Fine-Tuning) | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 16.6 | 1.1 |
| 学习不遗忘(Learning without Forgetting,LwF) [144] | 0.1 | 0.0 | 0.4 | 2.6 | 4.6 | 16.9 | 1.7 |
| 多分类学习不遗忘(Learning without Forgetting for Multi-Class,LwF-MC) [216] | 18.7 | 2.5 | 8.7 | 4.1 | 6.5 | 5.1 | 14.3 |
| 增量式学习训练(Incremental Learning Training,ILT) [178] | 0.3 | 0.0 | 1.0 | 2.1 | 4.6 | 10.7 | 1.4 |
| 兆字节(Mebibyte,MiB) | 31.8 | 10.4 | 14.8 | 12.8 | 13.6 | 18.7 | 25.9 |
| 联合 | 44.3 | 26.1 | 42.8 | 26.7 | 28.1 | 17.3 | 38.9 |











| Method | KnownClasses | OWR | ||||
| 11 | 16 | 21 | 26 | [20] | H | |
| GC | 66.0 | 57.3 | 58.6 | 53.3 | 58.8 | 58.7 |
| LC | 64.1 | 56.0 | 57.9 | 56.4 | 58.6 | 58.4 |
| Triplet | 62.1 | 54.9 | 54.8 | 49.5 | 55.4 | 55.4 |
| 67.7 | 59.6 | 59.5 | 57.3 | 61.0 | 60.8 | |
| 方法 | 已知类别 | 开放世界识别(Open World Recognition,OWR) | ||||
| 11 | 16 | 21 | 26 | [20] | H | |
| 全局约束(Global Constraint,GC) | 66.0 | 57.3 | 58.6 | 53.3 | 58.8 | 58.7 |
| 局部约束(Local Constraint,LC) | 64.1 | 56.0 | 57.9 | 56.4 | 58.6 | 58.4 |
| 三元组 | 62.1 | 54.9 | 54.8 | 49.5 | 55.4 | 55.4 |
| 67.7 | 59.6 | 59.5 | 57.3 | 61.0 | 60.8 | |
| Method | Class specific | Known | Unknown | Diff. | |
| DeepNNO | 84.4 | 98.8 | 14.4 | ||
| B-DOC | ✓ | 83.0 | 98.6 | 15.6 | |
| ✓ | 4.4 | 26.9 | 22.6 | ||
| ✓ | ✓ | 27.4 | 65.2 | 37.8 |
| 方法 | 特定类别 | 已知 | 未知 | 差异 | |
| 深度神经网络优化器(DeepNNO) | 84.4 | 98.8 | 14.4 | ||
| 基于边界的离群点检测(B - DOC) | ✓ | 83.0 | 98.6 | 15.6 | |
| ✓ | 4.4 | 26.9 | 22.6 | ||
| ✓ | ✓ | 27.4 | 65.2 | 37.8 |







| Target | AGG | DANN [78] | MLDG [134] | CrossGrad [238] | MetaReg [10] | JiGen [27] | Epi-FCR [135] | CuMix |
| Photo | 94.9 | 94.0 | 94.3 | 94.0 | 94.3 | 96.0 | 93.9 | 95.1 |
| Art | 76.1 | 81.3 | 79.5 | 78.7 | 79.5 | 79.4 | 82.1 | 82.3 |
| Cartoon | 73.8 | 73.8 | 77.3 | 73.3 | 75.4 | 75.3 | 77.0 | 76.5 |
| Sketch | 69.4 | 74.3 | 71.5 | 65.1 | 72.2 | 71.4 | 73.0 | 72.6 |
| Average | 78.5 | 80.8 | 80.7 | 80.7 | 77.8 | 80.4 | 81.5 | 81.6 |
| 目标 | 聚合(AGG) | 领域对抗神经网络(DANN) [78] | 多任务学习领域泛化(MLDG) [134] | 交叉梯度(CrossGrad) [238] | 元正则化(MetaReg) [10] | 拼图生成(JiGen) [27] | 情景式特征对比正则化(Epi - FCR) [135] | 混合裁剪(CuMix) |
| 照片 | 94.9 | 94.0 | 94.3 | 94.0 | 94.3 | 96.0 | 93.9 | 95.1 |
| 艺术画 | 76.1 | 81.3 | 79.5 | 78.7 | 79.5 | 79.4 | 82.1 | 82.3 |
| 卡通画 | 73.8 | 73.8 | 77.3 | 73.3 | 75.4 | 75.3 | 77.0 | 76.5 |
| 素描 | 69.4 | 74.3 | 71.5 | 65.1 | 72.2 | 71.4 | 73.0 | 72.6 |
| 平均值 | 78.5 | 80.8 | 80.7 | 80.7 | 77.8 | 80.4 | 81.5 | 81.6 |
| Curriculum | Art | Cartoon | Photo | Sketch | Avg. | |||
| ✓ | 76.1 | 73.8 | 94.9 | 69.4 | 78.5 | |||
| ✓ | ✓ | 78.4 | 72.7 | 94.7 | 59.5 | 76.3 | ||
| ✓ | ✓ | 81.8 | 76.5 | 94.9 | 71.2 | 81.1 | ||
| ✓ | ✓ | ✓ | 82.7 | 75.4 | 95.4 | 71.5 | 81.2 | |
| ✓ | ✓ | ✓ | ✓ | 82.3 | 76.5 | 95.1 | 72.6 | 81.6 |
| 课程表(Curriculum) | 艺术(Art) | 卡通(Cartoon) | 照片(Photo) | 素描(Sketch) | 平均(Avg.) | |||
| ✓ | 76.1 | 73.8 | 94.9 | 69.4 | 78.5 | |||
| ✓ | ✓ | 78.4 | 72.7 | 94.7 | 59.5 | 76.3 | ||
| ✓ | ✓ | 81.8 | 76.5 | 94.9 | 71.2 | 81.1 | ||
| ✓ | ✓ | ✓ | 82.7 | 75.4 | 95.4 | 71.5 | 81.2 | |
| ✓ | ✓ | ✓ | ✓ | 82.3 | 76.5 | 95.1 | 72.6 | 81.6 |
| Method | Clipart | Infograph | Painting | Quickdraw | Sketch | Avg. |
| SPNet | 26.0 | 16.9 | 23.8 | 8.2 | 21.8 | 19.4 |
| mixup+SPNet | 27.2 | 16.9 | 24.7 | 8.5 | 21.3 | 19.7 |
| Epi-FCR+SPNet | 26.4 | 16.7 | 24.6 | 9.2 | 23.2 | 20.0 |
| CuMix | 27.6 | 17.8 | 25.5 | 9.9 | 22.6 | 20.7 |
| 方法 | 剪贴画 | 信息图 | 绘画 | 快速绘图 | 素描 | 平均值 |
| SP网络(SPNet) | 26.0 | 16.9 | 23.8 | 8.2 | 21.8 | 19.4 |
| 混合增强+SP网络(mixup+SPNet) | 27.2 | 16.9 | 24.7 | 8.5 | 21.3 | 19.7 |
| 表型特征对比正则化+SP网络(Epi-FCR+SPNet) | 26.4 | 16.7 | 24.6 | 9.2 | 23.2 | 20.0 |
| CuMix(暂未找到通用译法,保留原文) | 27.6 | 17.8 | 25.5 | 9.9 | 22.6 | 20.7 |



| Method | Sketch | Photo | Art | Cartoon | Mean |
| ResNet [98] | 60.1 | 92.9 | 74.7 | 72.4 | 75.0 |
| DIAL [29] | 66.8 (71.3) | 97.0 (97.4) | 87.3 (87.5) | 85.5 (87.0) | 84.2 (85.8) |
| mDA | 70.7 (75.2) | 97.0 (97.3) | 87.4 (87.7) | 86.3 (87.2) | 85.4 (86.9) |
| Multi-source DA | 71.6 (78.1) | 96.6 (97.2) | 87.5 (88.7) | 87.0 (87.4) | 85.7 (87.9) |
| 方法 | 草图 | 照片 | 艺术 | 卡通 | 均值 |
| 残差网络(ResNet) [98] | 60.1 | 92.9 | 74.7 | 72.4 | 75.0 |
| DIAL [29] | 66.8 (71.3) | 97.0 (97.4) | 87.3 (87.5) | 85.5 (87.0) | 84.2 (85.8) |
| 多重去噪自编码器(mDA) | 70.7 (75.2) | 97.0 (97.3) | 87.4 (87.7) | 86.3 (87.2) | 85.4 (86.9) |
| 多源领域自适应(Multi - source DA) | 71.6 (78.1) | 96.6 (97.2) | 87.5 (88.7) | 87.0 (87.4) | 85.7 (87.9) |
| Method | Photo-Art | Photo-Cartoon | Photo-Sketch | Art-Cartoon | Art-Sketch | Cartoon-Sketch | Mean |
| ResNet [98] | 71.4 | 84.2 | 81.4 | 62.2 | 70.3 | 54.2 | 70.6 |
| DIAL [29] | 86.7 (87.5) | 86.5 (87.1) | 77.1 (78.7) | 72.1 (74.2) | 67.7 (70.4) | 79.5 (81.0) | |
| mDA | 87.2 (87.7) | 88.1 (88.5) | 88.7 (89.7) | 77.7 (79.6) | 81.3 (82.2) | 77.0 (79.3) | 83.3 (84.5) |
| Multi-source/ target DA | 79.0 (79.5) | 79.8 (82.2) | 75.6 (79.1) |
| 方法 | 照片艺术(Photo-Art) | 照片卡通(Photo-Cartoon) | 照片素描(Photo-Sketch) | 艺术卡通(Art-Cartoon) | 艺术素描(Art-Sketch) | 卡通素描(Cartoon-Sketch) | 均值 |
| 残差网络(ResNet [98]) | 71.4 | 84.2 | 81.4 | 62.2 | 70.3 | 54.2 | 70.6 |
| DIAL [29](原文未明确,保留英文) | 86.7 (87.5) | 86.5 (87.1) | 77.1 (78.7) | 72.1 (74.2) | 67.7 (70.4) | 79.5 (81.0) | |
| 多去噪自编码器(mDA,原文未明确,保留英文) | 87.2 (87.7) | 88.1 (88.5) | 88.7 (89.7) | 77.7 (79.6) | 81.3 (82.2) | 77.0 (79.3) | 83.3 (84.5) |
| 多源/目标域适应(Multi-source/ target DA) | 79.0 (79.5) | 79.8 (82.2) | 75.6 (79.1) |
| Method | Avg. Accuracy |
| Baseline | 56.8 |
| AdaGraph | 65.1 |
| Baseline + Refinement | 65.3 |
| AdaGraph + Refinement | 66.7 |
| DA upper bound | 66.9 |
| 方法 | 平均准确率 |
| 基线(Baseline) | 56.8 |
| 自适应图(AdaGraph) | 65.1 |
| 基线 + 细化 | 65.3 |
| 自适应图 + 细化 | 66.7 |
| 领域自适应上限(DA upper bound) | 66.9 |

| 19-1 | 15-5 | 15-1 | |||||||
| Method | 1-19 | 20 | all | 1-15 | 16-20 | all | 1-15 | 16-20 | |
| LwF-MC-C | 44.6 | 17.6 | 43.2 | 41.6 | 42.2 | 41.8 | 4.4 | 8.6 | 5.4 |
| LwF-MC | 64.4 | 13.3 | 61.9 | 58.1 | 35.0 | 52.3 | 6.4 | 8.4 | 6.9 |
| LwF-MC-D | 71.3 | 3.6 | 68.0 | 73.7 | 21.0 | 60.5 | 41.1 | 3.1 | 31.6 |
| MiB | 70.2 | 22.1 | 67.8 | 75.5 | 49.4 | 69.0 | 35.1 | 13.5 | 29.7 |
| Joint | 77.4 | 78.0 | 77.4 | 79.1 | 72.6 | 77.4 | 79.1 | 72.6 | 77.4 |
| 19-1 | 15-5 | 15-1 | |||||||
| 方法 | 1-19 | 20 | 全部 | 1-15 | 16-20 | 全部 | 1-15 | 16-20 | |
| 带多分类约束的学习遗忘法(LwF-MC-C) | 44.6 | 17.6 | 43.2 | 41.6 | 42.2 | 41.8 | 4.4 | 8.6 | 5.4 |
| 多分类学习遗忘法(LwF-MC) | 64.4 | 13.3 | 61.9 | 58.1 | 35.0 | 52.3 | 6.4 | 8.4 | 6.9 |
| 带多分类蒸馏的学习遗忘法(LwF-MC-D) | 71.3 | 3.6 | 68.0 | 73.7 | 21.0 | 60.5 | 41.1 | 3.1 | 31.6 |
| 兆字节(MiB) | 70.2 | 22.1 | 67.8 | 75.5 | 49.4 | 69.0 | 35.1 | 13.5 | 29.7 |
| 联合 | 77.4 | 78.0 | 77.4 | 79.1 | 72.6 | 77.4 | 79.1 | 72.6 | 77.4 |
| Method | 1-50 | 51-100 | 101-150 | all |
| LwF-MC-C | 8.0 | 7.2 | 19.3 | 11.5 |
| LwF-MC | 27.8 | 7.0 | 10.4 | 15.1 |
| LwF-MC-D | 39.1 | 10.9 | 6.7 | 18.7 |
| MiB | 35.5 | 22.2 | 23.6 | 27.0 |
| Joint | 51.1 | 38.3 | 28.2 | 38.9 |
| 方法 | 1-50 | 51-100 | 101-150 | 全部 |
| 带多分类约束的学习无遗忘方法(LwF-MC-C) | 8.0 | 7.2 | 19.3 | 11.5 |
| 多分类学习无遗忘方法(LwF-MC) | 27.8 | 7.0 | 10.4 | 15.1 |
| 带多分类蒸馏的学习无遗忘方法(LwF-MC-D) | 39.1 | 10.9 | 6.7 | 18.7 |
| 兆字节(MiB) | 35.5 | 22.2 | 23.6 | 27.0 |
| 联合 | 51.1 | 38.3 | 28.2 | 38.9 |
| Method | aero | bike | bird | boat | bottle | bus | car | cat | chair | COW | table | dog | horse | mbike | persn | plant | sheep | sofa | train | tv | 1-19 | all |
| FT | 11.9 | 2.1 | 1.1 | 11.6 | 4.8 | 6.9 | 13.5 | 0.2 | 0.0 | 3.8 | 14.4 | 0.5 | 1.5 | 4.7 | 0.0 | 15.8 | 2.8 | 1.8 | 13.5 | 12.3 | 5.8 | 6.2 |
| PI [300] | 22.3 | 1.9 | 3.4 | 4.9 | 2.1 | 10.6 | 8.5 | 0.1 | 0.1 | 3.1 | 12.8 | 0.2 | 3.8 | 4.6 | 0.0 | 10.0 | 5.0 | 1.1 | 8.5 | 14.1 | 5.4 | 5.9 |
| EWC [118] | 50.7 | 7.7 | 21.0 | 24.1 | 21.8 | 35.8 | 43.9 | 11.6 | 2.0 | 27.0 | 21.1 | 23.0 | 18.7 | 19.4 | 1.5 | 27.8 | 41.5 | 5.6 | 37.4 | 16.0 | 23.2 | 22.9 |
| RW [36] | 45.8 | 5.3 | 15.1 | 22.8 | 17.8 | 28.9 | 40.9 | 7.5 | 1.3 | 22.4 | 20.3 | 14.5 | 13.7 | 16.3 | 0.8 | 25.3 | 31.8 | 4.8 | 33.3 | 15.7 | 19.4 | 19.2 |
| LwF [144] | 28.1 | 40.5 | 53.1 | 38.8 | 47.4 | 46.4 | 63.6 | 83.5 | 35.8 | 60.1 | 48.8 | 76.5 | 65.3 | 67.1 | 83.2 | 50.2 | 61.2 | 42.5 | 14.2 | 9.1 | 53.0 | 50.8 |
| LwF-MC [216] | 79.4 | 41.3 | 75.6 | 47.9 | 51.0 | 69.6 | 75.4 | 78.5 | 35.1 | 66.6 | 49.0 | 72.7 | 73.8 | 71.6 | 84.9 | 57.5 | 67.7 | 42.7 | 56.8 | 13.2 | 63.0 | 60.5 |
| ILT [178] | 83.7 | 40.8 | 80.8 | 59.1 | 58.4 | 77.6 | 82.4 | 82.3 | 38.9 | 81.7 | 50.8 | 84.8 | 86.6 | 81.0 | 83.3 | 56.4 | 82.2 | 43.8 | 57.5 | 16.4 | 69.1 | 66.4 |
| MiB | 78.0 | 40.5 | 85.7 | 51.6 | 64.4 | 79.1 | 77.8 | 89.9 | 39.2 | 82.3 | 55.4 | 86.2 | 82.7 | 72.2 | 83.6 | 56.6 | 86.2 | 45.1 | 65.0 | 25.6 | 69.6 | 67.4 |
| Joint | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 77.4 | 77.4 |
| 方法 | 航空(aero) | 自行车 | 鸟 | 船 | 瓶子 | 公共汽车 | 汽车 | 猫 | 椅子 | 奶牛 | 桌子 | 狗 | 马 | 摩托车(mbike) | 人(persn) | 植物 | 绵羊 | 沙发 | 火车 | 电视 | 1-19 | 全部 |
| FT(未明确,保留英文) | 11.9 | 2.1 | 1.1 | 11.6 | 4.8 | 6.9 | 13.5 | 0.2 | 0.0 | 3.8 | 14.4 | 0.5 | 1.5 | 4.7 | 0.0 | 15.8 | 2.8 | 1.8 | 13.5 | 12.3 | 5.8 | 6.2 |
| PI [300](未明确,保留英文) | 22.3 | 1.9 | 3.4 | 4.9 | 2.1 | 10.6 | 8.5 | 0.1 | 0.1 | 3.1 | 12.8 | 0.2 | 3.8 | 4.6 | 0.0 | 10.0 | 5.0 | 1.1 | 8.5 | 14.1 | 5.4 | 5.9 |
| EWC [118](未明确,保留英文) | 50.7 | 7.7 | 21.0 | 24.1 | 21.8 | 35.8 | 43.9 | 11.6 | 2.0 | 27.0 | 21.1 | 23.0 | 18.7 | 19.4 | 1.5 | 27.8 | 41.5 | 5.6 | 37.4 | 16.0 | 23.2 | 22.9 |
| RW [36](未明确,保留英文) | 45.8 | 5.3 | 15.1 | 22.8 | 17.8 | 28.9 | 40.9 | 7.5 | 1.3 | 22.4 | 20.3 | 14.5 | 13.7 | 16.3 | 0.8 | 25.3 | 31.8 | 4.8 | 33.3 | 15.7 | 19.4 | 19.2 |
| LwF [144](未明确,保留英文) | 28.1 | 40.5 | 53.1 | 38.8 | 47.4 | 46.4 | 63.6 | 83.5 | 35.8 | 60.1 | 48.8 | 76.5 | 65.3 | 67.1 | 83.2 | 50.2 | 61.2 | 42.5 | 14.2 | 9.1 | 53.0 | 50.8 |
| LwF - MC [216](未明确,保留英文) | 79.4 | 41.3 | 75.6 | 47.9 | 51.0 | 69.6 | 75.4 | 78.5 | 35.1 | 66.6 | 49.0 | 72.7 | 73.8 | 71.6 | 84.9 | 57.5 | 67.7 | 42.7 | 56.8 | 13.2 | 63.0 | 60.5 |
| ILT [178](未明确,保留英文) | 83.7 | 40.8 | 80.8 | 59.1 | 58.4 | 77.6 | 82.4 | 82.3 | 38.9 | 81.7 | 50.8 | 84.8 | 86.6 | 81.0 | 83.3 | 56.4 | 82.2 | 43.8 | 57.5 | 16.4 | 69.1 | 66.4 |
| 兆字节(MiB) | 78.0 | 40.5 | 85.7 | 51.6 | 64.4 | 79.1 | 77.8 | 89.9 | 39.2 | 82.3 | 55.4 | 86.2 | 82.7 | 72.2 | 83.6 | 56.6 | 86.2 | 45.1 | 65.0 | 25.6 | 69.6 | 67.4 |
| 联合 | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 77.4 | 77.4 |
| Method | aero | bike | bird | boat | bottle | bus | car | cat | chair | COW | table | dog | horse | mbike | persn | plant | sheep | sofa | train | tv | 1-19 | all |
| FT | 23.7 | 1.9 | 1.5 | 9.3 | 6.9 | 16.9 | 8.5 | 0.0 | 0.0 | 9.5 | 5.3 | 0.1 | 2.9 | 8.8 | 0.0 | 15.1 | 1.0 | 0.7 | 16.0 | 12.9 | 6.8 | 7.1 |
| PI [300] | 33.1 | 4.1 | 3.6 | 10.5 | 8.4 | 14.7 | 13.3 | 0.0 | 0.1 | 2.4 | 4.7 | 0.1 | 3.3 | 7.9 | 0.0 | 14.7 | 0.8 | 2.7 | 17.8 | 14.0 | 7.5 | 7.8 |
| EWC [118] | 60.7 | 14.8 | 21.2 | 33.8 | 36.9 | 54.4 | 45.6 | 2.6 | 1.4 | 33.0 | 13.3 | 19.1 | 23.8 | 39.2 | 2.2 | 34.6 | 21.8 | 6.4 | 47.1 | 14.0 | 26.9 | 26.3 |
| RW [36] | 57.5 | 12.1 | 15.4 | 29.6 | 32.9 | 50.7 | 40.0 | 1.3 | 0.8 | 30.7 | 10.7 | 12.6 | 18.6 | 32.9 | 0.8 | 30.7 | 17.5 | 5.5 | 42.7 | 14.2 | 23.3 | 22.9 |
| LwF [144] | 36.6 | 35.1 | 62.0 | 32.9 | 47.5 | 31.6 | 51.5 | 77.9 | 36.5 | 67.7 | 44.3 | 71.4 | 68.6 | 66.2 | 82.2 | 49.6 | 58.7 | 41.1 | 11.9 | 8.5 | 51.2 | 49.1 |
| LwF-MC [216] | 67.2 | 37.9 | 77.8 | 40.6 | 57.0 | 54.5 | 77.4 | 88.4 | 37.2 | 76.8 | 49.1 | 83.4 | 82.3 | 71.0 | 85.2 | 55.6 | 81.9 | 46.0 | 54.9 | 13.3 | 64.4 | 61.9 |
| ILT [178] | 87.2 | 39.0 | 80.6 | 53.5 | 57.0 | 80.3 | 76.0 | 74.3 | 37.6 | 81.1 | 44.6 | 83.1 | 84.4 | 81.6 | 82.4 | 54.5 | 82.7 | 38.9 | 56.1 | 12.3 | 67.1 | 64.4 |
| MiB | 78.1 | 36.2 | 86.8 | 49.4 | 72.7 | 80.8 | 78.2 | 90.8 | 38.3 | 82.0 | 51.9 | 86.7 | 82.8 | 76.9 | 83.8 | 58.8 | 84.4 | 45.7 | 68.5 | 22.1 | 70.2 | 67.8 |
| Joint | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 77.4 | 77.4 |
| 方法 | 航空(aero) | 自行车 | 鸟 | 船 | 瓶子 | 公共汽车 | 汽车 | 猫 | 椅子 | 奶牛 | 桌子 | 狗 | 马 | 摩托车(mbike) | 人(persn) | 植物 | 绵羊 | 沙发 | 火车 | 电视 | 1-19 | 全部 |
| FT(未明确,保留英文) | 23.7 | 1.9 | 1.5 | 9.3 | 6.9 | 16.9 | 8.5 | 0.0 | 0.0 | 9.5 | 5.3 | 0.1 | 2.9 | 8.8 | 0.0 | 15.1 | 1.0 | 0.7 | 16.0 | 12.9 | 6.8 | 7.1 |
| PI [300](未明确,保留英文) | 33.1 | 4.1 | 3.6 | 10.5 | 8.4 | 14.7 | 13.3 | 0.0 | 0.1 | 2.4 | 4.7 | 0.1 | 3.3 | 7.9 | 0.0 | 14.7 | 0.8 | 2.7 | 17.8 | 14.0 | 7.5 | 7.8 |
| EWC [118](未明确,保留英文) | 60.7 | 14.8 | 21.2 | 33.8 | 36.9 | 54.4 | 45.6 | 2.6 | 1.4 | 33.0 | 13.3 | 19.1 | 23.8 | 39.2 | 2.2 | 34.6 | 21.8 | 6.4 | 47.1 | 14.0 | 26.9 | 26.3 |
| RW [36](未明确,保留英文) | 57.5 | 12.1 | 15.4 | 29.6 | 32.9 | 50.7 | 40.0 | 1.3 | 0.8 | 30.7 | 10.7 | 12.6 | 18.6 | 32.9 | 0.8 | 30.7 | 17.5 | 5.5 | 42.7 | 14.2 | 23.3 | 22.9 |
| LwF [144](未明确,保留英文) | 36.6 | 35.1 | 62.0 | 32.9 | 47.5 | 31.6 | 51.5 | 77.9 | 36.5 | 67.7 | 44.3 | 71.4 | 68.6 | 66.2 | 82.2 | 49.6 | 58.7 | 41.1 | 11.9 | 8.5 | 51.2 | 49.1 |
| LwF - MC [216](未明确,保留英文) | 67.2 | 37.9 | 77.8 | 40.6 | 57.0 | 54.5 | 77.4 | 88.4 | 37.2 | 76.8 | 49.1 | 83.4 | 82.3 | 71.0 | 85.2 | 55.6 | 81.9 | 46.0 | 54.9 | 13.3 | 64.4 | 61.9 |
| ILT [178](未明确,保留英文) | 87.2 | 39.0 | 80.6 | 53.5 | 57.0 | 80.3 | 76.0 | 74.3 | 37.6 | 81.1 | 44.6 | 83.1 | 84.4 | 81.6 | 82.4 | 54.5 | 82.7 | 38.9 | 56.1 | 12.3 | 67.1 | 64.4 |
| 兆字节(MiB) | 78.1 | 36.2 | 86.8 | 49.4 | 72.7 | 80.8 | 78.2 | 90.8 | 38.3 | 82.0 | 51.9 | 86.7 | 82.8 | 76.9 | 83.8 | 58.8 | 84.4 | 45.7 | 68.5 | 22.1 | 70.2 | 67.8 |
| 联合;共同 | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 77.4 | 77.4 |
| Method | aero | bike | bird | boat | bottle | bus | car | cat | chair | COW | table | dog | horse | mbike | persn | plant | sheep | sofa | train | tv | 1-15 | 16-20 | all |
| FT | 6.1 | 0.0 | 0.2 | 8.3 | 0.1 | 0.0 | 0.1 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 1.8 | 0.0 | 0.0 | 24.6 | 24.3 | 36.2 | 32.5 | 50.2 | 1.1 | 33.6 | 9.2 |
| PI [300] | 8.8 | 0.0 | 0.2 | 10.5 | 0.0 | 0.0 | 0.1 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.4 | 0.0 | 0.0 | 25.6 | 24.7 | 34.3 | 34.1 | 52.0 | 1.3 | 34.1 | 9.5 |
| EWC [118] | 58.8 | 4.1 | 56.4 | 46.2 | 44.4 | 4.3 | 67.4 | 3.6 | 2.3 | 14.8 | 10.3 | 12.4 | 51.6 | 20.4 | 2.9 | 28.8 | 32.2 | 35.6 | 35.5 | 56.3 | 26.7 | 37.7 | 29.4 |
| RW [36] | 51.1 | 1.5 | 36.9 | 42.9 | 27.5 | 2.1 | 47.4 | 1.1 | 1.2 | 6.1 | 5.3 | 3.1 | 31.2 | 10.5 | 1.0 | 27.7 | 29.8 | 35.7 | 34.7 | 56.6 | 17.9 | 36.9 | 22.7 |
| LwF [144] | 63.1 | 40.1 | 72.4 | 52.1 | 67.0 | 6.7 | 80.3 | 84.2 | 31.1 | 5.7 | 51.3 | 82.0 | 75.0 | 79.4 | 85.6 | 35.3 | 27.1 | 37.0 | 37.0 | 50.5 | 58.4 | 37.4 | 53.1 |
| LwF-MC [216] | 78.1 | 42.3 | 78.9 | 62.1 | 78.6 | 47.3 | 84.6 | 89.1 | 35.0 | 26.2 | 50.5 | 86.6 | 77.6 | 84.9 | 86.0 | 35.0 | 35.2 | 40.8 | 49.2 | 45.9 | 67.2 | 41.2 | 60.7 |
| ILT [178] | 79.4 | 42.0 | 80.5 | 63.9 | 80.4 | 12.8 | 86.0 | 90.2 | 30.7 | 6.7 | 53.3 | 83.2 | 73.0 | 80.7 | 85.0 | 36.9 | 29.9 | 36.8 | 38.3 | 55.7 | 63.2 | 39.5 | 57.3 |
| MiB | 84.4 | 39.4 | 87.5 | 65.2 | 77.8 | 61.0 | 86.0 | 90.9 | 35.3 | 60.3 | 53.0 | 88.2 | 80.4 | 82.4 | 85.3 | 28.7 | 46.0 | 34.7 | 54.4 | 52.7 | 71.8 | 43.3 | 64.7 |
| Joint | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 79.1 | 72.6 | 77.4 |
| 方法 | 航空(aero) | 自行车 | 鸟 | 船 | 瓶子 | 公共汽车 | 汽车 | 猫 | 椅子 | 奶牛 | 桌子 | 狗 | 马 | 摩托车(mbike) | 人(persn) | 植物 | 绵羊 | 沙发 | 火车 | 电视 | 1-15 | 16-20 | 全部 |
| FT(未明确,保留英文) | 6.1 | 0.0 | 0.2 | 8.3 | 0.1 | 0.0 | 0.1 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 1.8 | 0.0 | 0.0 | 24.6 | 24.3 | 36.2 | 32.5 | 50.2 | 1.1 | 33.6 | 9.2 |
| PI [300](未明确,保留英文) | 8.8 | 0.0 | 0.2 | 10.5 | 0.0 | 0.0 | 0.1 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.4 | 0.0 | 0.0 | 25.6 | 24.7 | 34.3 | 34.1 | 52.0 | 1.3 | 34.1 | 9.5 |
| EWC [118](未明确,保留英文) | 58.8 | 4.1 | 56.4 | 46.2 | 44.4 | 4.3 | 67.4 | 3.6 | 2.3 | 14.8 | 10.3 | 12.4 | 51.6 | 20.4 | 2.9 | 28.8 | 32.2 | 35.6 | 35.5 | 56.3 | 26.7 | 37.7 | 29.4 |
| RW [36](未明确,保留英文) | 51.1 | 1.5 | 36.9 | 42.9 | 27.5 | 2.1 | 47.4 | 1.1 | 1.2 | 6.1 | 5.3 | 3.1 | 31.2 | 10.5 | 1.0 | 27.7 | 29.8 | 35.7 | 34.7 | 56.6 | 17.9 | 36.9 | 22.7 |
| LwF [144](未明确,保留英文) | 63.1 | 40.1 | 72.4 | 52.1 | 67.0 | 6.7 | 80.3 | 84.2 | 31.1 | 5.7 | 51.3 | 82.0 | 75.0 | 79.4 | 85.6 | 35.3 | 27.1 | 37.0 | 37.0 | 50.5 | 58.4 | 37.4 | 53.1 |
| LwF - MC [216](未明确,保留英文) | 78.1 | 42.3 | 78.9 | 62.1 | 78.6 | 47.3 | 84.6 | 89.1 | 35.0 | 26.2 | 50.5 | 86.6 | 77.6 | 84.9 | 86.0 | 35.0 | 35.2 | 40.8 | 49.2 | 45.9 | 67.2 | 41.2 | 60.7 |
| ILT [178](未明确,保留英文) | 79.4 | 42.0 | 80.5 | 63.9 | 80.4 | 12.8 | 86.0 | 90.2 | 30.7 | 6.7 | 53.3 | 83.2 | 73.0 | 80.7 | 85.0 | 36.9 | 29.9 | 36.8 | 38.3 | 55.7 | 63.2 | 39.5 | 57.3 |
| 兆字节(MiB) | 84.4 | 39.4 | 87.5 | 65.2 | 77.8 | 61.0 | 86.0 | 90.9 | 35.3 | 60.3 | 53.0 | 88.2 | 80.4 | 82.4 | 85.3 | 28.7 | 46.0 | 34.7 | 54.4 | 52.7 | 71.8 | 43.3 | 64.7 |
| 联合 | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 79.1 | 72.6 | 77.4 |
| Method | aero | bike | bird | boat | bottle | bus | car | cat | chair | COW | table | dog | horse | mbike | persn | plant | sheep | sofa | train | tv | 1-15 | 16-20 | all |
| FT | 13.4 | 0.1 | 0.0 | 15.6 | 0.8 | 0.0 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.9 | 0.0 | 0.0 | 30.9 | 21.6 | 32.8 | 34.9 | 45.1 | 2.1 | 33.1 | 9.8 |
| PI [300] | 7.8 | 0.0 | 0.0 | 12.9 | 0.3 | 0.0 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 2.7 | 0.0 | 0.0 | 33.2 | 22.2 | 33.2 | 36.1 | 42.0 | 1.6 | 33.3 | 9.5 |
| EWC [118] | 67.3 | 12.8 | 50.5 | 52.9 | 35.0 | 24.7 | 41.7 | 1.2 | 1.0 | 9.8 | 5.7 | 3.7 | 42.9 | 15.4 | 0.6 | 31.8 | 26.3 | 32.1 | 42.0 | 45.0 | 24.3 | 35.5 | 27.1 |
| RW [36] | 61.2 | 6.7 | 33.8 | 48.1 | 24.4 | 9.3 | 22.3 | 0.3 | 0.5 | 3.5 | 0.2 | 1.1 | 31.8 | 6.4 | 0.1 | 32.1 | 25.8 | 31.9 | 38.7 | 45.9 | 16.6 | 34.9 | 21.2 |
| LwF [144] | 64.5 | 40.2 | 72.8 | 56.9 | 57.3 | 9.5 | 82.6 | 88.6 | 33.2 | 8.9 | 48.4 | 81.9 | 75.0 | 78.2 | 84.9 | 34.7 | 27.8 | 33.1 | 39.6 | 48.0 | 58.9 | 36.6 | 53.3 |
| LwF-MC [216] | 60.6 | 38.9 | 74.7 | 41.6 | 67.2 | 10.8 | 81.4 | 88.8 | 38.7 | 4.3 | 47.4 | 82.2 | 69.9 | 78.9 | 85.8 | 28.4 | 28.5 | 34.1 | 36.4 | 47.8 | 58.1 | 35.0 | 52.3 |
| ILT [178] | 77.4 | 40.3 | 78.9 | 61.9 | 78.7 | 53.5 | 86.1 | 88.7 | 33.8 | 15.9 | 51.1 | 83.2 | 80.2 | 79.8 | 85.0 | 39.5 | 30.9 | 31.0 | 49.3 | 52.6 | 66.3 | 40.6 | 59.9 |
| MiB | 86.6 | 39.3 | 88.9 | 66.1 | 80.8 | 86.6 | 90.1 | 92.5 | 38.0 | 64.6 | 56.4 | 89.6 | 80.5 | 86.5 | 85.7 | 30.2 | 52.9 | 31.3 | 73.2 | 59.5 | 75.5 | 49.4 | 69.0 |
| Joint | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 79.1 | 72.6 | 77.4 |
| 方法 | 航空(aero) | 自行车 | 鸟 | 船 | 瓶子 | 公共汽车 | 汽车 | 猫 | 椅子 | 奶牛 | 桌子 | 狗 | 马 | 摩托车(mbike) | 人(persn) | 植物 | 绵羊 | 沙发 | 火车 | 电视 | 1-15 | 16-20 | 全部 |
| FT(未明确,保留英文) | 13.4 | 0.1 | 0.0 | 15.6 | 0.8 | 0.0 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.9 | 0.0 | 0.0 | 30.9 | 21.6 | 32.8 | 34.9 | 45.1 | 2.1 | 33.1 | 9.8 |
| PI [300](未明确,保留英文) | 7.8 | 0.0 | 0.0 | 12.9 | 0.3 | 0.0 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 2.7 | 0.0 | 0.0 | 33.2 | 22.2 | 33.2 | 36.1 | 42.0 | 1.6 | 33.3 | 9.5 |
| EWC [118](未明确,保留英文) | 67.3 | 12.8 | 50.5 | 52.9 | 35.0 | 24.7 | 41.7 | 1.2 | 1.0 | 9.8 | 5.7 | 3.7 | 42.9 | 15.4 | 0.6 | 31.8 | 26.3 | 32.1 | 42.0 | 45.0 | 24.3 | 35.5 | 27.1 |
| RW [36](未明确,保留英文) | 61.2 | 6.7 | 33.8 | 48.1 | 24.4 | 9.3 | 22.3 | 0.3 | 0.5 | 3.5 | 0.2 | 1.1 | 31.8 | 6.4 | 0.1 | 32.1 | 25.8 | 31.9 | 38.7 | 45.9 | 16.6 | 34.9 | 21.2 |
| LwF [144](未明确,保留英文) | 64.5 | 40.2 | 72.8 | 56.9 | 57.3 | 9.5 | 82.6 | 88.6 | 33.2 | 8.9 | 48.4 | 81.9 | 75.0 | 78.2 | 84.9 | 34.7 | 27.8 | 33.1 | 39.6 | 48.0 | 58.9 | 36.6 | 53.3 |
| LwF - MC [216](未明确,保留英文) | 60.6 | 38.9 | 74.7 | 41.6 | 67.2 | 10.8 | 81.4 | 88.8 | 38.7 | 4.3 | 47.4 | 82.2 | 69.9 | 78.9 | 85.8 | 28.4 | 28.5 | 34.1 | 36.4 | 47.8 | 58.1 | 35.0 | 52.3 |
| ILT [178](未明确,保留英文) | 77.4 | 40.3 | 78.9 | 61.9 | 78.7 | 53.5 | 86.1 | 88.7 | 33.8 | 15.9 | 51.1 | 83.2 | 80.2 | 79.8 | 85.0 | 39.5 | 30.9 | 31.0 | 49.3 | 52.6 | 66.3 | 40.6 | 59.9 |
| 兆字节(MiB) | 86.6 | 39.3 | 88.9 | 66.1 | 80.8 | 86.6 | 90.1 | 92.5 | 38.0 | 64.6 | 56.4 | 89.6 | 80.5 | 86.5 | 85.7 | 30.2 | 52.9 | 31.3 | 73.2 | 59.5 | 75.5 | 49.4 | 69.0 |
| 联合 | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 79.1 | 72.6 | 77.4 |
| Method | aero | bike | bird | boat | bottle | bus | car | cat | chair | COW | table | dog | horse | mbike | persn | plant | sheep | sofa | train | tv | 1-15 | 16-20 | all |
| FT | 0.3 | 0.0 | 0.0 | 2.5 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 8.8 | 0.2 | 1.8 | 0.6 |
| PI [300] | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.3 | 8.6 | 0.0 | 1.8 | 0.4 |
| EWC [118] | 0.0 | 0.0 | 0.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 3.6 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 7.3 | 7.0 | 7.4 | 0.3 | 4.3 | 1.3 |
| RW [36] | 0.0 | 0.0 | 0.0 | 0.2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 2.2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 8.1 | 10.5 | 8.2 | 0.2 | 5.4 | 1.5 |
| LwF [144] | 0.0 | 0.0 | 0.0 | 0.0 | 0.6 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 10.7 | 0.0 | 0.0 | 1.9 | 8.2 | 7.9 | 0.8 | 3.6 | 1.5 |
| LwF-MC [216] | 0.0 | 6.3 | 0.8 | 0.0 | 1.1 | 0.0 | 0.1 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.2 | 0.0 | 59.0 | 0.0 | 9.5 | 2.9 | 11.9 | 11.0 | 4.5 | 7.0 | 5.2 |
| ILT [178] | 3.7 | 0.0 | 2.9 | 0.0 | 12.8 | 0.0 | 0.0 | 0.1 | 0.0 | 0.0 | 21.2 | 0.1 | 0.4 | 0.6 | 13.6 | 0.0 | 0.0 | 11.6 | 8.3 | 8.5 | 3.7 | 5.7 | 4.2 |
| MiB | 53.6 | 38.9 | 53.6 | 17.7 | 62.7 | 36.5 | 71.2 | 60.1 | 1.1 | 35.2 | 8.1 | 57.6 | 55.0 | 62.1 | 79.4 | 10.2 | 14.2 | 11.9 | 18.2 | 10.1 | 46.2 | 12.9 | 37.9 |
| Joint | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 79.1 | 72.6 | 77.4 |
| 方法 | 航空(aero) | 自行车 | 鸟 | 船 | 瓶子 | 公共汽车 | 汽车 | 猫 | 椅子 | 奶牛 | 桌子 | 狗 | 马 | 摩托车(mbike) | 人(persn) | 植物 | 绵羊 | 沙发 | 火车 | 电视 | 1-15 | 16-20 | 全部 |
| FT(未明确,保留英文) | 0.3 | 0.0 | 0.0 | 2.5 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 8.8 | 0.2 | 1.8 | 0.6 |
| PI [300](未明确,保留英文) | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.3 | 8.6 | 0.0 | 1.8 | 0.4 |
| EWC [118](未明确,保留英文) | 0.0 | 0.0 | 0.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 3.6 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 7.3 | 7.0 | 7.4 | 0.3 | 4.3 | 1.3 |
| RW [36](未明确,保留英文) | 0.0 | 0.0 | 0.0 | 0.2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 2.2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 8.1 | 10.5 | 8.2 | 0.2 | 5.4 | 1.5 |
| LwF [144](未明确,保留英文) | 0.0 | 0.0 | 0.0 | 0.0 | 0.6 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 10.7 | 0.0 | 0.0 | 1.9 | 8.2 | 7.9 | 0.8 | 3.6 | 1.5 |
| LwF - MC [216](未明确,保留英文) | 0.0 | 6.3 | 0.8 | 0.0 | 1.1 | 0.0 | 0.1 | 0.3 | 0.0 | 0.0 | 0.0 | 0.0 | 0.2 | 0.0 | 59.0 | 0.0 | 9.5 | 2.9 | 11.9 | 11.0 | 4.5 | 7.0 | 5.2 |
| ILT [178](未明确,保留英文) | 3.7 | 0.0 | 2.9 | 0.0 | 12.8 | 0.0 | 0.0 | 0.1 | 0.0 | 0.0 | 21.2 | 0.1 | 0.4 | 0.6 | 13.6 | 0.0 | 0.0 | 11.6 | 8.3 | 8.5 | 3.7 | 5.7 | 4.2 |
| 兆字节(MiB) | 53.6 | 38.9 | 53.6 | 17.7 | 62.7 | 36.5 | 71.2 | 60.1 | 1.1 | 35.2 | 8.1 | 57.6 | 55.0 | 62.1 | 79.4 | 10.2 | 14.2 | 11.9 | 18.2 | 10.1 | 46.2 | 12.9 | 37.9 |
| 联合 | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 79.1 | 72.6 | 77.4 |
| Method | aero | bike | bird | boat | bottle | bus | car | cat | chair | cow | table | dog | horse | mbike | persn | plant | sheep | sofa | train | tv | 1-15 | 16-20 | all |
| FT | 2.6 | 0.0 | 0.0 | 0.7 | 0.0 | 0.1 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 9.2 | 0.2 | 1.8 | 0.6 |
| PI [300] | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.2 | 9.1 | 0.0 | 1.8 | 0.5 |
| EWC [118] | 0.0 | 0.0 | 0.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 3.6 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 7.3 | 7.0 | 7.4 | 0.3 | 4.3 | 1.3 |
| RW [36] | 0.1 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 8.7 | 11.2 | 6.3 | 0.0 | 5.2 | 1.3 |
| LwF [144] | 3.7 | 0.1 | 0.0 | 2.5 | 0.2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.0 | 9.0 | 0.0 | 0.0 | 1.6 | 8.9 | 8.8 | 1.0 | 3.9 | 1.8 |
| LwF-MC [216] | 0.0 | 7.2 | 5.2 | 0.0 | 25.5 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 1.2 | 1.3 | 56.2 | 0.0 | 4.9 | 0.2 | 8.6 | 28.2 | 6.4 | 8.4 | 6.9 |
| ILT [178] | 20.0 | 0.0 | 3.2 | 6.3 | 2.3 | 0.0 | 0.0 | 0.0 | 0.3 | 5.1 | 19.0 | 0.0 | 9.1 | 0.0 | 8.7 | 0.0 | 0.0 | 21.0 | 9.9 | 8.1 | 4.9 | 7.8 | 5.7 |
| MiB | 31.3 | 25.4 | 26.7 | 26.9 | 46.1 | 31.0 | 63.6 | 52.8 | 0.1 | 11.0 | 9.4 | 52.4 | 41.2 | 28.1 | 80.7 | 17.6 | 13.1 | 15.3 | 15.3 | 6.2 | 35.1 | 13.5 | 29.7 |
| Joint | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 79.1 | 72.6 | 77.4 |
| 方法 | 航空(aero) | 自行车 | 鸟 | 船 | 瓶子 | 公共汽车 | 汽车 | 猫 | 椅子 | 奶牛 | 桌子 | 狗 | 马 | 摩托车(mbike) | 人(persn) | 植物 | 绵羊 | 沙发 | 火车 | 电视 | 1-15 | 16-20 | 全部 |
| FT(未明确,保留英文) | 2.6 | 0.0 | 0.0 | 0.7 | 0.0 | 0.1 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 9.2 | 0.2 | 1.8 | 0.6 |
| PI [300](未明确,保留英文) | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.2 | 9.1 | 0.0 | 1.8 | 0.5 |
| EWC [118](未明确,保留英文) | 0.0 | 0.0 | 0.0 | 1.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 3.6 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 7.3 | 7.0 | 7.4 | 0.3 | 4.3 | 1.3 |
| RW [36](未明确,保留英文) | 0.1 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 8.7 | 11.2 | 6.3 | 0.0 | 5.2 | 1.3 |
| LwF [144](未明确,保留英文) | 3.7 | 0.1 | 0.0 | 2.5 | 0.2 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.1 | 0.0 | 9.0 | 0.0 | 0.0 | 1.6 | 8.9 | 8.8 | 1.0 | 3.9 | 1.8 |
| LwF - MC [216](未明确,保留英文) | 0.0 | 7.2 | 5.2 | 0.0 | 25.5 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 1.2 | 1.3 | 56.2 | 0.0 | 4.9 | 0.2 | 8.6 | 28.2 | 6.4 | 8.4 | 6.9 |
| ILT [178](未明确,保留英文) | 20.0 | 0.0 | 3.2 | 6.3 | 2.3 | 0.0 | 0.0 | 0.0 | 0.3 | 5.1 | 19.0 | 0.0 | 9.1 | 0.0 | 8.7 | 0.0 | 0.0 | 21.0 | 9.9 | 8.1 | 4.9 | 7.8 | 5.7 |
| 兆字节(MiB) | 31.3 | 25.4 | 26.7 | 26.9 | 46.1 | 31.0 | 63.6 | 52.8 | 0.1 | 11.0 | 9.4 | 52.4 | 41.2 | 28.1 | 80.7 | 17.6 | 13.1 | 15.3 | 15.3 | 6.2 | 35.1 | 13.5 | 29.7 |
| 联合 | 90.2 | 42.2 | 89.5 | 69.1 | 82.3 | 92.5 | 90.0 | 94.2 | 39.2 | 87.6 | 56.4 | 91.2 | 86.8 | 88.0 | 86.8 | 62.3 | 88.4 | 49.5 | 85.0 | 78.0 | 79.1 | 72.6 | 77.4 |
| Method | Target Domain | ||||||
| DG | ZSL | clipart | infograph | painting | quickdraw | sketch | avg. |
| - | DEVISE [73] | 20.1 | 11.7 | 17.6 | 6.1 | 16.7 | 14.4 |
| ALE [1] | 22.7 | 12.7 | 20.2 | 6.8 | 18.5 | 16.2 | |
| SPNet [277] | 26.0 | 16.9 | 23.8 | 8.2 | 21.8 | 19.4 | |
| DANN [78] | DEVISE [73] | 20.5 | 10.4 | 16.4 | 7.1 | 15.1 | 13.9 |
| ALE [1] | 21.2 | 12.5 | 19.7 | 7.4 | 17.9 | 15.7 | |
| SPNet [277] | 25.9 | 15.8 | 24.1 | 8.4 | 21.3 | 19.1 | |
| EpiFCR [135] | DEVISE [73] | 21.6 | 13.9 | 19.3 | 7.3 | 17.2 | 15.9 |
| ALE [1] | 23.2 | 14.1 | 21.4 | 7.8 | 20.9 | 17.5 | |
| SPNet [277] | 26.4 | 16.7 | 24.6 | 9.2 | 23.2 | 20.0 | |
| CuMix | 27.6 | 17.8 | 25.5 | 9.9 | 22.6 | 20.7 | |
| 方法 | 目标领域 | ||||||
| DG(域泛化,Domain Generalization) | ZSL(零样本学习,Zero-Shot Learning) | 剪贴画 | 信息图 | 绘画 | 快速绘图 | 素描 | 平均值 |
| - | DEVISE [73](原文未明确,保留英文) | 20.1 | 11.7 | 17.6 | 6.1 | 16.7 | 14.4 |
| ALE [1](原文未明确,保留英文) | 22.7 | 12.7 | 20.2 | 6.8 | 18.5 | 16.2 | |
| SPNet [277](原文未明确,保留英文) | 26.0 | 16.9 | 23.8 | 8.2 | 21.8 | 19.4 | |
| DANN [78](原文未明确,保留英文) | DEVISE [73](原文未明确,保留英文) | 20.5 | 10.4 | 16.4 | 7.1 | 15.1 | 13.9 |
| ALE [1](原文未明确,保留英文) | 21.2 | 12.5 | 19.7 | 7.4 | 17.9 | 15.7 | |
| SPNet [277](原文未明确,保留英文) | 25.9 | 15.8 | 24.1 | 8.4 | 21.3 | 19.1 | |
| EpiFCR [135](原文未明确,保留英文) | DEVISE [73](原文未明确,保留英文) | 21.6 | 13.9 | 19.3 | 7.3 | 17.2 | 15.9 |
| ALE [1](原文未明确,保留英文) | 23.2 | 14.1 | 21.4 | 7.8 | 20.9 | 17.5 | |
| SPNet [277](原文未明确,保留英文) | 26.4 | 16.7 | 24.6 | 9.2 | 23.2 | 20.0 | |
| CuMix(原文未明确,保留英文) | 27.6 | 17.8 | 25.5 | 9.9 | 22.6 | 20.7 | |
| Method/Target | Clipart | Infograph | Sketch | Quickdraw | Avg. |
| SPNet | 14.4 | ||||
| Epi-FCR+SPNet | 15.4 | ||||
| MixUp img only | 14.3 | ||||
| MixUp two-level | 15.8 | ||||
| CuMix reverse | 15.4 | ||||
| CuMix | 17.1±0.2 | 16.5 |
| 方法/目标 | 剪贴画 | 信息图 | 草图 | 快速绘图 | 平均值 |
| SP网络 | 14.4 | ||||
| 外延式全卷积回归网络+SP网络 | 15.4 | ||||
| 仅图像混合增强 | 14.3 | ||||
| 两级混合增强 | 15.8 | ||||
| 反向CuMix | 15.4 | ||||
| CuMix | 17.1±0.2 | 16.5 |
| Method | CUB | SUN | AWA1 | FLO |
| ALE [1] | 54.9 | 58.1 | 59.9 | 48.5 |
| SJE [2] | 53.9 | 53.7 | 65.6 | 53.4 |
| SYNC [34] | 56.3 | 55.6 | 54.0 | - |
| GFZSL [265] | 49.3 | 60.6 | 68.3 | - |
| SPNet [277] | 56.5 | 60.7 | 66.2 | - |
| Baseline | 52.4 | 58.2 | 62.5 | 58.4 |
| CuMix | 60.4 | 62.4 | 64.0 | 59.7 |
| 方法 | 加州理工学院鸟类数据集(CUB) | 斯坦福大学场景数据集(SUN) | 动物属性数据集1(AWA1) | 花卉数据集(FLO) |
| ALE [1] | 54.9 | 58.1 | 59.9 | 48.5 |
| SJE [2] | 53.9 | 53.7 | 65.6 | 53.4 |
| SYNC [34] | 56.3 | 55.6 | 54.0 | - |
| 广义零样本学习方法(GFZSL) [265] | 49.3 | 60.6 | 68.3 | - |
| SPNet [277] | 56.5 | 60.7 | 66.2 | - |
| 基线 | 52.4 | 58.2 | 62.5 | 58.4 |
| CuMix | 60.4 | 62.4 | 64.0 | 59.7 |